Model save
Browse files- README.md +7 -5
- all_results.json +8 -0
- generation_config.json +14 -0
- model-00001-of-00004.safetensors +1 -1
- model-00002-of-00004.safetensors +1 -1
- model-00003-of-00004.safetensors +1 -1
- model-00004-of-00004.safetensors +1 -1
- train_results.json +8 -0
- trainer_state.json +2856 -0
README.md
CHANGED
|
@@ -1,31 +1,33 @@
|
|
| 1 |
---
|
| 2 |
base_model: eyad-silx/Quasar-3.0-Max
|
| 3 |
library_name: transformers
|
| 4 |
-
model_name: Quasar-3.
|
| 5 |
tags:
|
| 6 |
- generated_from_trainer
|
| 7 |
- trl
|
| 8 |
- sft
|
| 9 |
licence: license
|
| 10 |
---
|
| 11 |
-
# Model Card for Quasar-3.3-Max
|
| 12 |
|
| 13 |
-
|
|
|
|
|
|
|
| 14 |
It has been trained using [TRL](https://github.com/huggingface/trl).
|
| 15 |
|
| 16 |
## Quick start
|
| 17 |
|
| 18 |
```python
|
| 19 |
from transformers import pipeline
|
|
|
|
| 20 |
question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
|
| 21 |
-
generator = pipeline("text-generation", model="silx-ai/Quasar-3.
|
| 22 |
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
|
| 23 |
print(output["generated_text"])
|
| 24 |
```
|
| 25 |
|
| 26 |
## Training procedure
|
| 27 |
|
| 28 |
-
[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/mentoxcompuny-silx-ai/huggingface/runs/
|
| 29 |
|
| 30 |
|
| 31 |
This model was trained with SFT.
|
|
|
|
| 1 |
---
|
| 2 |
base_model: eyad-silx/Quasar-3.0-Max
|
| 3 |
library_name: transformers
|
| 4 |
+
model_name: Quasar-3.0-Max
|
| 5 |
tags:
|
| 6 |
- generated_from_trainer
|
| 7 |
- trl
|
| 8 |
- sft
|
| 9 |
licence: license
|
| 10 |
---
|
|
|
|
| 11 |
|
| 12 |
+
# Model Card for Quasar-3.0-Max
|
| 13 |
+
|
| 14 |
+
This model is a fine-tuned version of [eyad-silx/Quasar-3.0-Max](https://huggingface.co/eyad-silx/Quasar-3.0-Max).
|
| 15 |
It has been trained using [TRL](https://github.com/huggingface/trl).
|
| 16 |
|
| 17 |
## Quick start
|
| 18 |
|
| 19 |
```python
|
| 20 |
from transformers import pipeline
|
| 21 |
+
|
| 22 |
question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
|
| 23 |
+
generator = pipeline("text-generation", model="silx-ai/Quasar-3.0-Max", device="cuda")
|
| 24 |
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
|
| 25 |
print(output["generated_text"])
|
| 26 |
```
|
| 27 |
|
| 28 |
## Training procedure
|
| 29 |
|
| 30 |
+
[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/mentoxcompuny-silx-ai/huggingface/runs/wmstn4dn)
|
| 31 |
|
| 32 |
|
| 33 |
This model was trained with SFT.
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"total_flos": 8.574354709534474e+17,
|
| 3 |
+
"train_loss": 0.3579157315083404,
|
| 4 |
+
"train_runtime": 11555.4492,
|
| 5 |
+
"train_samples": 28743,
|
| 6 |
+
"train_samples_per_second": 6.688,
|
| 7 |
+
"train_steps_per_second": 0.035
|
| 8 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token_id": 151643,
|
| 3 |
+
"do_sample": true,
|
| 4 |
+
"eos_token_id": [
|
| 5 |
+
151645,
|
| 6 |
+
151643
|
| 7 |
+
],
|
| 8 |
+
"pad_token_id": 151643,
|
| 9 |
+
"repetition_penalty": 1.05,
|
| 10 |
+
"temperature": 0.7,
|
| 11 |
+
"top_k": 20,
|
| 12 |
+
"top_p": 0.8,
|
| 13 |
+
"transformers_version": "4.49.0"
|
| 14 |
+
}
|
model-00001-of-00004.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4877660776
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b989467fcc81931d91f5e64ca44815fff79c16b05640dce6a4c3df31a12c8171
|
| 3 |
size 4877660776
|
model-00002-of-00004.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4932751008
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:12f564ae74512f091921bc427ce5011f6550f6b214ea1666e51f2297afd41e6f
|
| 3 |
size 4932751008
|
model-00003-of-00004.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4330865200
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d7aad5317a48a6bb60c5802cf7644ccad0026ac11857a47b28b572a78f7a1ed4
|
| 3 |
size 4330865200
|
model-00004-of-00004.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1089994880
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:37c522af9365989e77ce4f6e5e06165c40201a43e2d33efce4a7d9005d3c3e4e
|
| 3 |
size 1089994880
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"total_flos": 8.574354709534474e+17,
|
| 3 |
+
"train_loss": 0.3579157315083404,
|
| 4 |
+
"train_runtime": 11555.4492,
|
| 5 |
+
"train_samples": 28743,
|
| 6 |
+
"train_samples_per_second": 6.688,
|
| 7 |
+
"train_steps_per_second": 0.035
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2856 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.9838509316770185,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 402,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.007453416149068323,
|
| 13 |
+
"grad_norm": 0.9149440636536349,
|
| 14 |
+
"learning_rate": 2.3809523809523808e-06,
|
| 15 |
+
"loss": 0.5267,
|
| 16 |
+
"step": 1
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.014906832298136646,
|
| 20 |
+
"grad_norm": 0.813081054678093,
|
| 21 |
+
"learning_rate": 4.7619047619047615e-06,
|
| 22 |
+
"loss": 0.501,
|
| 23 |
+
"step": 2
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.02236024844720497,
|
| 27 |
+
"grad_norm": 0.7030352003534567,
|
| 28 |
+
"learning_rate": 7.142857142857143e-06,
|
| 29 |
+
"loss": 0.4786,
|
| 30 |
+
"step": 3
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.02981366459627329,
|
| 34 |
+
"grad_norm": 0.744974289676234,
|
| 35 |
+
"learning_rate": 9.523809523809523e-06,
|
| 36 |
+
"loss": 0.5096,
|
| 37 |
+
"step": 4
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.037267080745341616,
|
| 41 |
+
"grad_norm": 0.5298082392046644,
|
| 42 |
+
"learning_rate": 1.1904761904761905e-05,
|
| 43 |
+
"loss": 0.4947,
|
| 44 |
+
"step": 5
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.04472049689440994,
|
| 48 |
+
"grad_norm": 0.6276792078680746,
|
| 49 |
+
"learning_rate": 1.4285714285714285e-05,
|
| 50 |
+
"loss": 0.4889,
|
| 51 |
+
"step": 6
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.05217391304347826,
|
| 55 |
+
"grad_norm": 0.9938211412587247,
|
| 56 |
+
"learning_rate": 1.6666666666666667e-05,
|
| 57 |
+
"loss": 0.4966,
|
| 58 |
+
"step": 7
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.05962732919254658,
|
| 62 |
+
"grad_norm": 1.0688888961766796,
|
| 63 |
+
"learning_rate": 1.9047619047619046e-05,
|
| 64 |
+
"loss": 0.5423,
|
| 65 |
+
"step": 8
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.0670807453416149,
|
| 69 |
+
"grad_norm": 0.7660208255317018,
|
| 70 |
+
"learning_rate": 2.1428571428571428e-05,
|
| 71 |
+
"loss": 0.5054,
|
| 72 |
+
"step": 9
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.07453416149068323,
|
| 76 |
+
"grad_norm": 0.6232752556675165,
|
| 77 |
+
"learning_rate": 2.380952380952381e-05,
|
| 78 |
+
"loss": 0.516,
|
| 79 |
+
"step": 10
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.08198757763975155,
|
| 83 |
+
"grad_norm": 0.5696191527866266,
|
| 84 |
+
"learning_rate": 2.6190476190476192e-05,
|
| 85 |
+
"loss": 0.4902,
|
| 86 |
+
"step": 11
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.08944099378881988,
|
| 90 |
+
"grad_norm": 0.5574577130188229,
|
| 91 |
+
"learning_rate": 2.857142857142857e-05,
|
| 92 |
+
"loss": 0.5054,
|
| 93 |
+
"step": 12
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.0968944099378882,
|
| 97 |
+
"grad_norm": 0.5992840639020023,
|
| 98 |
+
"learning_rate": 3.095238095238095e-05,
|
| 99 |
+
"loss": 0.5008,
|
| 100 |
+
"step": 13
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.10434782608695652,
|
| 104 |
+
"grad_norm": 0.5954549641006874,
|
| 105 |
+
"learning_rate": 3.3333333333333335e-05,
|
| 106 |
+
"loss": 0.5023,
|
| 107 |
+
"step": 14
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.11180124223602485,
|
| 111 |
+
"grad_norm": 0.576921385529578,
|
| 112 |
+
"learning_rate": 3.571428571428572e-05,
|
| 113 |
+
"loss": 0.5236,
|
| 114 |
+
"step": 15
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.11925465838509317,
|
| 118 |
+
"grad_norm": 0.5706806482628634,
|
| 119 |
+
"learning_rate": 3.809523809523809e-05,
|
| 120 |
+
"loss": 0.5017,
|
| 121 |
+
"step": 16
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.1267080745341615,
|
| 125 |
+
"grad_norm": 0.7086057812964052,
|
| 126 |
+
"learning_rate": 4.047619047619048e-05,
|
| 127 |
+
"loss": 0.5267,
|
| 128 |
+
"step": 17
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.1341614906832298,
|
| 132 |
+
"grad_norm": 0.5808757241964403,
|
| 133 |
+
"learning_rate": 4.2857142857142856e-05,
|
| 134 |
+
"loss": 0.4904,
|
| 135 |
+
"step": 18
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.14161490683229813,
|
| 139 |
+
"grad_norm": 0.6317506464382385,
|
| 140 |
+
"learning_rate": 4.523809523809524e-05,
|
| 141 |
+
"loss": 0.5043,
|
| 142 |
+
"step": 19
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.14906832298136646,
|
| 146 |
+
"grad_norm": 0.6359839242618298,
|
| 147 |
+
"learning_rate": 4.761904761904762e-05,
|
| 148 |
+
"loss": 0.5055,
|
| 149 |
+
"step": 20
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.1565217391304348,
|
| 153 |
+
"grad_norm": 0.5752065827486615,
|
| 154 |
+
"learning_rate": 5e-05,
|
| 155 |
+
"loss": 0.5084,
|
| 156 |
+
"step": 21
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.1639751552795031,
|
| 160 |
+
"grad_norm": 0.574109201187312,
|
| 161 |
+
"learning_rate": 4.999923510846293e-05,
|
| 162 |
+
"loss": 0.4981,
|
| 163 |
+
"step": 22
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.17142857142857143,
|
| 167 |
+
"grad_norm": 0.5996134853952727,
|
| 168 |
+
"learning_rate": 4.999694048585699e-05,
|
| 169 |
+
"loss": 0.4889,
|
| 170 |
+
"step": 23
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.17888198757763976,
|
| 174 |
+
"grad_norm": 0.6523805926616552,
|
| 175 |
+
"learning_rate": 4.999311628819437e-05,
|
| 176 |
+
"loss": 0.4927,
|
| 177 |
+
"step": 24
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.18633540372670807,
|
| 181 |
+
"grad_norm": 0.7101608078052835,
|
| 182 |
+
"learning_rate": 4.9987762775483666e-05,
|
| 183 |
+
"loss": 0.5078,
|
| 184 |
+
"step": 25
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.1937888198757764,
|
| 188 |
+
"grad_norm": 0.7396307593577484,
|
| 189 |
+
"learning_rate": 4.99808803117121e-05,
|
| 190 |
+
"loss": 0.4943,
|
| 191 |
+
"step": 26
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.20124223602484473,
|
| 195 |
+
"grad_norm": 0.653880275554884,
|
| 196 |
+
"learning_rate": 4.9972469364820877e-05,
|
| 197 |
+
"loss": 0.4981,
|
| 198 |
+
"step": 27
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.20869565217391303,
|
| 202 |
+
"grad_norm": 0.9242693837414091,
|
| 203 |
+
"learning_rate": 4.996253050667329e-05,
|
| 204 |
+
"loss": 0.4875,
|
| 205 |
+
"step": 28
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.21614906832298136,
|
| 209 |
+
"grad_norm": 1.2095187282407587,
|
| 210 |
+
"learning_rate": 4.995106441301587e-05,
|
| 211 |
+
"loss": 0.5071,
|
| 212 |
+
"step": 29
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.2236024844720497,
|
| 216 |
+
"grad_norm": 0.7405451123886392,
|
| 217 |
+
"learning_rate": 4.993807186343243e-05,
|
| 218 |
+
"loss": 0.5168,
|
| 219 |
+
"step": 30
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.231055900621118,
|
| 223 |
+
"grad_norm": 0.6770762633940498,
|
| 224 |
+
"learning_rate": 4.992355374129109e-05,
|
| 225 |
+
"loss": 0.4956,
|
| 226 |
+
"step": 31
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.23850931677018633,
|
| 230 |
+
"grad_norm": 0.8790648152676741,
|
| 231 |
+
"learning_rate": 4.990751103368418e-05,
|
| 232 |
+
"loss": 0.4914,
|
| 233 |
+
"step": 32
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.24596273291925466,
|
| 237 |
+
"grad_norm": 0.8761927581585033,
|
| 238 |
+
"learning_rate": 4.988994483136115e-05,
|
| 239 |
+
"loss": 0.5028,
|
| 240 |
+
"step": 33
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.253416149068323,
|
| 244 |
+
"grad_norm": 1.0155828765501556,
|
| 245 |
+
"learning_rate": 4.98708563286544e-05,
|
| 246 |
+
"loss": 0.497,
|
| 247 |
+
"step": 34
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.2608695652173913,
|
| 251 |
+
"grad_norm": 0.7164731439245322,
|
| 252 |
+
"learning_rate": 4.985024682339807e-05,
|
| 253 |
+
"loss": 0.4782,
|
| 254 |
+
"step": 35
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.2683229813664596,
|
| 258 |
+
"grad_norm": 0.7789231779550446,
|
| 259 |
+
"learning_rate": 4.982811771683982e-05,
|
| 260 |
+
"loss": 0.5117,
|
| 261 |
+
"step": 36
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.27577639751552796,
|
| 265 |
+
"grad_norm": 0.7333852855548396,
|
| 266 |
+
"learning_rate": 4.980447051354555e-05,
|
| 267 |
+
"loss": 0.5147,
|
| 268 |
+
"step": 37
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.28322981366459626,
|
| 272 |
+
"grad_norm": 0.6356599304692622,
|
| 273 |
+
"learning_rate": 4.977930682129711e-05,
|
| 274 |
+
"loss": 0.4986,
|
| 275 |
+
"step": 38
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.2906832298136646,
|
| 279 |
+
"grad_norm": 0.6005754618090398,
|
| 280 |
+
"learning_rate": 4.975262835098295e-05,
|
| 281 |
+
"loss": 0.4853,
|
| 282 |
+
"step": 39
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.2981366459627329,
|
| 286 |
+
"grad_norm": 0.6269613620581922,
|
| 287 |
+
"learning_rate": 4.9724436916481836e-05,
|
| 288 |
+
"loss": 0.4842,
|
| 289 |
+
"step": 40
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.30559006211180123,
|
| 293 |
+
"grad_norm": 0.4930016834035119,
|
| 294 |
+
"learning_rate": 4.969473443453953e-05,
|
| 295 |
+
"loss": 0.5001,
|
| 296 |
+
"step": 41
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.3130434782608696,
|
| 300 |
+
"grad_norm": 0.6087046909213417,
|
| 301 |
+
"learning_rate": 4.966352292463842e-05,
|
| 302 |
+
"loss": 0.4903,
|
| 303 |
+
"step": 42
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.3204968944099379,
|
| 307 |
+
"grad_norm": 0.6680768802723248,
|
| 308 |
+
"learning_rate": 4.96308045088603e-05,
|
| 309 |
+
"loss": 0.4864,
|
| 310 |
+
"step": 43
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.3279503105590062,
|
| 314 |
+
"grad_norm": 0.5024028870492778,
|
| 315 |
+
"learning_rate": 4.959658141174196e-05,
|
| 316 |
+
"loss": 0.4928,
|
| 317 |
+
"step": 44
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.33540372670807456,
|
| 321 |
+
"grad_norm": 0.5691045158483018,
|
| 322 |
+
"learning_rate": 4.956085596012407e-05,
|
| 323 |
+
"loss": 0.4895,
|
| 324 |
+
"step": 45
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.34285714285714286,
|
| 328 |
+
"grad_norm": 0.5960753346949116,
|
| 329 |
+
"learning_rate": 4.952363058299291e-05,
|
| 330 |
+
"loss": 0.5037,
|
| 331 |
+
"step": 46
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.35031055900621116,
|
| 335 |
+
"grad_norm": 0.5809783536316744,
|
| 336 |
+
"learning_rate": 4.948490781131523e-05,
|
| 337 |
+
"loss": 0.4955,
|
| 338 |
+
"step": 47
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.3577639751552795,
|
| 342 |
+
"grad_norm": 0.4638121725312366,
|
| 343 |
+
"learning_rate": 4.944469027786616e-05,
|
| 344 |
+
"loss": 0.4828,
|
| 345 |
+
"step": 48
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.3652173913043478,
|
| 349 |
+
"grad_norm": 0.6519884038596414,
|
| 350 |
+
"learning_rate": 4.940298071705024e-05,
|
| 351 |
+
"loss": 0.4869,
|
| 352 |
+
"step": 49
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.37267080745341613,
|
| 356 |
+
"grad_norm": 0.725573010046233,
|
| 357 |
+
"learning_rate": 4.935978196471548e-05,
|
| 358 |
+
"loss": 0.4983,
|
| 359 |
+
"step": 50
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.3801242236024845,
|
| 363 |
+
"grad_norm": 0.7723414020223717,
|
| 364 |
+
"learning_rate": 4.931509695796055e-05,
|
| 365 |
+
"loss": 0.5097,
|
| 366 |
+
"step": 51
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.3875776397515528,
|
| 370 |
+
"grad_norm": 0.6749832378781871,
|
| 371 |
+
"learning_rate": 4.926892873493509e-05,
|
| 372 |
+
"loss": 0.5009,
|
| 373 |
+
"step": 52
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.3950310559006211,
|
| 377 |
+
"grad_norm": 0.5287344971699504,
|
| 378 |
+
"learning_rate": 4.922128043463316e-05,
|
| 379 |
+
"loss": 0.5033,
|
| 380 |
+
"step": 53
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.40248447204968946,
|
| 384 |
+
"grad_norm": 0.7590242284157045,
|
| 385 |
+
"learning_rate": 4.917215529667979e-05,
|
| 386 |
+
"loss": 0.5041,
|
| 387 |
+
"step": 54
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.40993788819875776,
|
| 391 |
+
"grad_norm": 0.9634027388127688,
|
| 392 |
+
"learning_rate": 4.9121556661110734e-05,
|
| 393 |
+
"loss": 0.5084,
|
| 394 |
+
"step": 55
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.41739130434782606,
|
| 398 |
+
"grad_norm": 0.8942287814733841,
|
| 399 |
+
"learning_rate": 4.906948796814539e-05,
|
| 400 |
+
"loss": 0.4982,
|
| 401 |
+
"step": 56
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.4248447204968944,
|
| 405 |
+
"grad_norm": 0.7332394854708344,
|
| 406 |
+
"learning_rate": 4.901595275795287e-05,
|
| 407 |
+
"loss": 0.4893,
|
| 408 |
+
"step": 57
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.4322981366459627,
|
| 412 |
+
"grad_norm": 0.6811910538390739,
|
| 413 |
+
"learning_rate": 4.896095467041133e-05,
|
| 414 |
+
"loss": 0.498,
|
| 415 |
+
"step": 58
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.43975155279503103,
|
| 419 |
+
"grad_norm": 1.0815817510677823,
|
| 420 |
+
"learning_rate": 4.8904497444860495e-05,
|
| 421 |
+
"loss": 0.4951,
|
| 422 |
+
"step": 59
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.4472049689440994,
|
| 426 |
+
"grad_norm": 0.5430186184806628,
|
| 427 |
+
"learning_rate": 4.884658491984735e-05,
|
| 428 |
+
"loss": 0.4974,
|
| 429 |
+
"step": 60
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.4546583850931677,
|
| 433 |
+
"grad_norm": 0.7264258373815644,
|
| 434 |
+
"learning_rate": 4.878722103286529e-05,
|
| 435 |
+
"loss": 0.5017,
|
| 436 |
+
"step": 61
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.462111801242236,
|
| 440 |
+
"grad_norm": 0.7414704944298521,
|
| 441 |
+
"learning_rate": 4.8726409820086274e-05,
|
| 442 |
+
"loss": 0.502,
|
| 443 |
+
"step": 62
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.46956521739130436,
|
| 447 |
+
"grad_norm": 0.523437918394872,
|
| 448 |
+
"learning_rate": 4.8664155416086495e-05,
|
| 449 |
+
"loss": 0.4957,
|
| 450 |
+
"step": 63
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.47701863354037266,
|
| 454 |
+
"grad_norm": 0.6159908506967338,
|
| 455 |
+
"learning_rate": 4.8600462053565204e-05,
|
| 456 |
+
"loss": 0.4956,
|
| 457 |
+
"step": 64
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.484472049689441,
|
| 461 |
+
"grad_norm": 0.5860033924334783,
|
| 462 |
+
"learning_rate": 4.8535334063056984e-05,
|
| 463 |
+
"loss": 0.5036,
|
| 464 |
+
"step": 65
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.4919254658385093,
|
| 468 |
+
"grad_norm": 0.510141066708458,
|
| 469 |
+
"learning_rate": 4.846877587263728e-05,
|
| 470 |
+
"loss": 0.4862,
|
| 471 |
+
"step": 66
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.4993788819875776,
|
| 475 |
+
"grad_norm": 0.5214414519844613,
|
| 476 |
+
"learning_rate": 4.840079200762135e-05,
|
| 477 |
+
"loss": 0.4728,
|
| 478 |
+
"step": 67
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.506832298136646,
|
| 482 |
+
"grad_norm": 0.5627288353125691,
|
| 483 |
+
"learning_rate": 4.833138709025656e-05,
|
| 484 |
+
"loss": 0.49,
|
| 485 |
+
"step": 68
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.5142857142857142,
|
| 489 |
+
"grad_norm": 0.6512461745013858,
|
| 490 |
+
"learning_rate": 4.826056583940815e-05,
|
| 491 |
+
"loss": 0.4763,
|
| 492 |
+
"step": 69
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.5217391304347826,
|
| 496 |
+
"grad_norm": 0.501857537261379,
|
| 497 |
+
"learning_rate": 4.818833307023839e-05,
|
| 498 |
+
"loss": 0.4801,
|
| 499 |
+
"step": 70
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.529192546583851,
|
| 503 |
+
"grad_norm": 0.5401175474298149,
|
| 504 |
+
"learning_rate": 4.811469369387917e-05,
|
| 505 |
+
"loss": 0.5099,
|
| 506 |
+
"step": 71
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.5366459627329192,
|
| 510 |
+
"grad_norm": 0.6056173499301607,
|
| 511 |
+
"learning_rate": 4.803965271709811e-05,
|
| 512 |
+
"loss": 0.4755,
|
| 513 |
+
"step": 72
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.5440993788819876,
|
| 517 |
+
"grad_norm": 0.5300538292613007,
|
| 518 |
+
"learning_rate": 4.7963215241958156e-05,
|
| 519 |
+
"loss": 0.4914,
|
| 520 |
+
"step": 73
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.5515527950310559,
|
| 524 |
+
"grad_norm": 0.49477384366828114,
|
| 525 |
+
"learning_rate": 4.7885386465470675e-05,
|
| 526 |
+
"loss": 0.4942,
|
| 527 |
+
"step": 74
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.5590062111801242,
|
| 531 |
+
"grad_norm": 0.5832215591212782,
|
| 532 |
+
"learning_rate": 4.780617167924209e-05,
|
| 533 |
+
"loss": 0.4821,
|
| 534 |
+
"step": 75
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.5664596273291925,
|
| 538 |
+
"grad_norm": 0.5825547831853878,
|
| 539 |
+
"learning_rate": 4.772557626911417e-05,
|
| 540 |
+
"loss": 0.4991,
|
| 541 |
+
"step": 76
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.5739130434782609,
|
| 545 |
+
"grad_norm": 0.6260775830743696,
|
| 546 |
+
"learning_rate": 4.7643605714797754e-05,
|
| 547 |
+
"loss": 0.5178,
|
| 548 |
+
"step": 77
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.5813664596273292,
|
| 552 |
+
"grad_norm": 0.5002047281833832,
|
| 553 |
+
"learning_rate": 4.756026558950025e-05,
|
| 554 |
+
"loss": 0.4788,
|
| 555 |
+
"step": 78
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.5888198757763975,
|
| 559 |
+
"grad_norm": 0.658164441734046,
|
| 560 |
+
"learning_rate": 4.747556155954668e-05,
|
| 561 |
+
"loss": 0.4924,
|
| 562 |
+
"step": 79
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.5962732919254659,
|
| 566 |
+
"grad_norm": 0.5642052228011414,
|
| 567 |
+
"learning_rate": 4.738949938399445e-05,
|
| 568 |
+
"loss": 0.4793,
|
| 569 |
+
"step": 80
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.6037267080745342,
|
| 573 |
+
"grad_norm": 0.6800799125862594,
|
| 574 |
+
"learning_rate": 4.730208491424174e-05,
|
| 575 |
+
"loss": 0.4972,
|
| 576 |
+
"step": 81
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.6111801242236025,
|
| 580 |
+
"grad_norm": 0.6022489226977471,
|
| 581 |
+
"learning_rate": 4.721332409362973e-05,
|
| 582 |
+
"loss": 0.4839,
|
| 583 |
+
"step": 82
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.6186335403726708,
|
| 587 |
+
"grad_norm": 0.5548484325825994,
|
| 588 |
+
"learning_rate": 4.712322295703846e-05,
|
| 589 |
+
"loss": 0.4786,
|
| 590 |
+
"step": 83
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.6260869565217392,
|
| 594 |
+
"grad_norm": 0.5717126912005164,
|
| 595 |
+
"learning_rate": 4.7031787630476544e-05,
|
| 596 |
+
"loss": 0.4938,
|
| 597 |
+
"step": 84
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.6335403726708074,
|
| 601 |
+
"grad_norm": 0.5272588715397946,
|
| 602 |
+
"learning_rate": 4.693902433066464e-05,
|
| 603 |
+
"loss": 0.4793,
|
| 604 |
+
"step": 85
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.6409937888198758,
|
| 608 |
+
"grad_norm": 0.5463413109564844,
|
| 609 |
+
"learning_rate": 4.6844939364612796e-05,
|
| 610 |
+
"loss": 0.4925,
|
| 611 |
+
"step": 86
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.6484472049689441,
|
| 615 |
+
"grad_norm": 0.6075766167098049,
|
| 616 |
+
"learning_rate": 4.674953912919161e-05,
|
| 617 |
+
"loss": 0.5025,
|
| 618 |
+
"step": 87
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.6559006211180124,
|
| 622 |
+
"grad_norm": 0.6310158980679005,
|
| 623 |
+
"learning_rate": 4.665283011069733e-05,
|
| 624 |
+
"loss": 0.4911,
|
| 625 |
+
"step": 88
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.6633540372670808,
|
| 629 |
+
"grad_norm": 0.50469934840197,
|
| 630 |
+
"learning_rate": 4.655481888441083e-05,
|
| 631 |
+
"loss": 0.4764,
|
| 632 |
+
"step": 89
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.6708074534161491,
|
| 636 |
+
"grad_norm": 0.488371832002747,
|
| 637 |
+
"learning_rate": 4.6455512114150546e-05,
|
| 638 |
+
"loss": 0.4937,
|
| 639 |
+
"step": 90
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.6782608695652174,
|
| 643 |
+
"grad_norm": 0.4770981226634187,
|
| 644 |
+
"learning_rate": 4.635491655181943e-05,
|
| 645 |
+
"loss": 0.4884,
|
| 646 |
+
"step": 91
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.6857142857142857,
|
| 650 |
+
"grad_norm": 0.5775420955722163,
|
| 651 |
+
"learning_rate": 4.625303903694586e-05,
|
| 652 |
+
"loss": 0.4853,
|
| 653 |
+
"step": 92
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.6931677018633541,
|
| 657 |
+
"grad_norm": 0.47430309169555146,
|
| 658 |
+
"learning_rate": 4.614988649621864e-05,
|
| 659 |
+
"loss": 0.5139,
|
| 660 |
+
"step": 93
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.7006211180124223,
|
| 664 |
+
"grad_norm": 0.6638176198018706,
|
| 665 |
+
"learning_rate": 4.604546594301602e-05,
|
| 666 |
+
"loss": 0.4953,
|
| 667 |
+
"step": 94
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.7080745341614907,
|
| 671 |
+
"grad_norm": 0.5170361990821142,
|
| 672 |
+
"learning_rate": 4.5939784476928894e-05,
|
| 673 |
+
"loss": 0.4714,
|
| 674 |
+
"step": 95
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.715527950310559,
|
| 678 |
+
"grad_norm": 0.5989923841003543,
|
| 679 |
+
"learning_rate": 4.583284928327805e-05,
|
| 680 |
+
"loss": 0.4826,
|
| 681 |
+
"step": 96
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.7229813664596273,
|
| 685 |
+
"grad_norm": 0.5119408388808395,
|
| 686 |
+
"learning_rate": 4.572466763262569e-05,
|
| 687 |
+
"loss": 0.4876,
|
| 688 |
+
"step": 97
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.7304347826086957,
|
| 692 |
+
"grad_norm": 0.4719741454775879,
|
| 693 |
+
"learning_rate": 4.5615246880281066e-05,
|
| 694 |
+
"loss": 0.4746,
|
| 695 |
+
"step": 98
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.737888198757764,
|
| 699 |
+
"grad_norm": 0.5792116907988438,
|
| 700 |
+
"learning_rate": 4.550459446580039e-05,
|
| 701 |
+
"loss": 0.4778,
|
| 702 |
+
"step": 99
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.7453416149068323,
|
| 706 |
+
"grad_norm": 0.5650326179874091,
|
| 707 |
+
"learning_rate": 4.5392717912481025e-05,
|
| 708 |
+
"loss": 0.5068,
|
| 709 |
+
"step": 100
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.7527950310559006,
|
| 713 |
+
"grad_norm": 0.49349361289777244,
|
| 714 |
+
"learning_rate": 4.527962482684998e-05,
|
| 715 |
+
"loss": 0.4787,
|
| 716 |
+
"step": 101
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.760248447204969,
|
| 720 |
+
"grad_norm": 0.590606894561532,
|
| 721 |
+
"learning_rate": 4.516532289814674e-05,
|
| 722 |
+
"loss": 0.4827,
|
| 723 |
+
"step": 102
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.7677018633540372,
|
| 727 |
+
"grad_norm": 0.6325065592193169,
|
| 728 |
+
"learning_rate": 4.5049819897800445e-05,
|
| 729 |
+
"loss": 0.4822,
|
| 730 |
+
"step": 103
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.7751552795031056,
|
| 734 |
+
"grad_norm": 0.48212618053402273,
|
| 735 |
+
"learning_rate": 4.493312367890154e-05,
|
| 736 |
+
"loss": 0.4862,
|
| 737 |
+
"step": 104
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.782608695652174,
|
| 741 |
+
"grad_norm": 0.6659769851227831,
|
| 742 |
+
"learning_rate": 4.481524217566783e-05,
|
| 743 |
+
"loss": 0.4763,
|
| 744 |
+
"step": 105
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.7900621118012422,
|
| 748 |
+
"grad_norm": 0.5100428848300065,
|
| 749 |
+
"learning_rate": 4.4696183402905005e-05,
|
| 750 |
+
"loss": 0.4861,
|
| 751 |
+
"step": 106
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.7975155279503106,
|
| 755 |
+
"grad_norm": 0.4930498732601481,
|
| 756 |
+
"learning_rate": 4.4575955455461764e-05,
|
| 757 |
+
"loss": 0.4709,
|
| 758 |
+
"step": 107
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.8049689440993789,
|
| 762 |
+
"grad_norm": 0.5557882232422525,
|
| 763 |
+
"learning_rate": 4.4454566507679395e-05,
|
| 764 |
+
"loss": 0.4823,
|
| 765 |
+
"step": 108
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.8124223602484472,
|
| 769 |
+
"grad_norm": 0.4779459811311798,
|
| 770 |
+
"learning_rate": 4.4332024812836026e-05,
|
| 771 |
+
"loss": 0.4902,
|
| 772 |
+
"step": 109
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.8198757763975155,
|
| 776 |
+
"grad_norm": 0.4951096303075202,
|
| 777 |
+
"learning_rate": 4.420833870258544e-05,
|
| 778 |
+
"loss": 0.4799,
|
| 779 |
+
"step": 110
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.8273291925465839,
|
| 783 |
+
"grad_norm": 0.4972554983396596,
|
| 784 |
+
"learning_rate": 4.4083516586390694e-05,
|
| 785 |
+
"loss": 0.4757,
|
| 786 |
+
"step": 111
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.8347826086956521,
|
| 790 |
+
"grad_norm": 0.6922157872179194,
|
| 791 |
+
"learning_rate": 4.395756695095222e-05,
|
| 792 |
+
"loss": 0.4871,
|
| 793 |
+
"step": 112
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.8422360248447205,
|
| 797 |
+
"grad_norm": 0.614533156193526,
|
| 798 |
+
"learning_rate": 4.383049835963095e-05,
|
| 799 |
+
"loss": 0.4858,
|
| 800 |
+
"step": 113
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.8496894409937888,
|
| 804 |
+
"grad_norm": 0.4998865414356172,
|
| 805 |
+
"learning_rate": 4.370231945186601e-05,
|
| 806 |
+
"loss": 0.5084,
|
| 807 |
+
"step": 114
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.8571428571428571,
|
| 811 |
+
"grad_norm": 0.5297858251800804,
|
| 812 |
+
"learning_rate": 4.357303894258733e-05,
|
| 813 |
+
"loss": 0.4844,
|
| 814 |
+
"step": 115
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.8645962732919255,
|
| 818 |
+
"grad_norm": 0.490173761162275,
|
| 819 |
+
"learning_rate": 4.344266562162313e-05,
|
| 820 |
+
"loss": 0.4775,
|
| 821 |
+
"step": 116
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.8720496894409938,
|
| 825 |
+
"grad_norm": 0.5188647432432411,
|
| 826 |
+
"learning_rate": 4.331120835310228e-05,
|
| 827 |
+
"loss": 0.4731,
|
| 828 |
+
"step": 117
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.8795031055900621,
|
| 832 |
+
"grad_norm": 0.5006594475547541,
|
| 833 |
+
"learning_rate": 4.3178676074851646e-05,
|
| 834 |
+
"loss": 0.4834,
|
| 835 |
+
"step": 118
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.8869565217391304,
|
| 839 |
+
"grad_norm": 0.6966182509694935,
|
| 840 |
+
"learning_rate": 4.3045077797788386e-05,
|
| 841 |
+
"loss": 0.482,
|
| 842 |
+
"step": 119
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.8944099378881988,
|
| 846 |
+
"grad_norm": 0.630693615513719,
|
| 847 |
+
"learning_rate": 4.29104226053073e-05,
|
| 848 |
+
"loss": 0.4806,
|
| 849 |
+
"step": 120
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.901863354037267,
|
| 853 |
+
"grad_norm": 0.4855641171083717,
|
| 854 |
+
"learning_rate": 4.277471965266325e-05,
|
| 855 |
+
"loss": 0.4881,
|
| 856 |
+
"step": 121
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.9093167701863354,
|
| 860 |
+
"grad_norm": 0.6046785410123138,
|
| 861 |
+
"learning_rate": 4.2637978166348666e-05,
|
| 862 |
+
"loss": 0.4985,
|
| 863 |
+
"step": 122
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.9167701863354037,
|
| 867 |
+
"grad_norm": 0.6094452228967235,
|
| 868 |
+
"learning_rate": 4.250020744346629e-05,
|
| 869 |
+
"loss": 0.488,
|
| 870 |
+
"step": 123
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.924223602484472,
|
| 874 |
+
"grad_norm": 0.5207025073876308,
|
| 875 |
+
"learning_rate": 4.2361416851097e-05,
|
| 876 |
+
"loss": 0.4839,
|
| 877 |
+
"step": 124
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.9316770186335404,
|
| 881 |
+
"grad_norm": 0.5910466009416144,
|
| 882 |
+
"learning_rate": 4.222161582566299e-05,
|
| 883 |
+
"loss": 0.4816,
|
| 884 |
+
"step": 125
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.9391304347826087,
|
| 888 |
+
"grad_norm": 0.6533302840950863,
|
| 889 |
+
"learning_rate": 4.208081387228612e-05,
|
| 890 |
+
"loss": 0.4848,
|
| 891 |
+
"step": 126
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.9465838509316771,
|
| 895 |
+
"grad_norm": 0.5615426924165942,
|
| 896 |
+
"learning_rate": 4.193902056414175e-05,
|
| 897 |
+
"loss": 0.4952,
|
| 898 |
+
"step": 127
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.9540372670807453,
|
| 902 |
+
"grad_norm": 0.5584154804823008,
|
| 903 |
+
"learning_rate": 4.179624554180778e-05,
|
| 904 |
+
"loss": 0.4998,
|
| 905 |
+
"step": 128
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.9614906832298137,
|
| 909 |
+
"grad_norm": 0.6235842837428364,
|
| 910 |
+
"learning_rate": 4.165249851260921e-05,
|
| 911 |
+
"loss": 0.4918,
|
| 912 |
+
"step": 129
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.968944099378882,
|
| 916 |
+
"grad_norm": 0.5109516402930713,
|
| 917 |
+
"learning_rate": 4.1507789249958134e-05,
|
| 918 |
+
"loss": 0.4812,
|
| 919 |
+
"step": 130
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.9763975155279503,
|
| 923 |
+
"grad_norm": 0.5909881028917956,
|
| 924 |
+
"learning_rate": 4.136212759268926e-05,
|
| 925 |
+
"loss": 0.4713,
|
| 926 |
+
"step": 131
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.9838509316770186,
|
| 930 |
+
"grad_norm": 0.8317014634966002,
|
| 931 |
+
"learning_rate": 4.121552344439093e-05,
|
| 932 |
+
"loss": 0.5024,
|
| 933 |
+
"step": 132
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.991304347826087,
|
| 937 |
+
"grad_norm": 0.49067977677281605,
|
| 938 |
+
"learning_rate": 4.1067986772731795e-05,
|
| 939 |
+
"loss": 0.4816,
|
| 940 |
+
"step": 133
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.9987577639751553,
|
| 944 |
+
"grad_norm": 0.6766338838361023,
|
| 945 |
+
"learning_rate": 4.0919527608783105e-05,
|
| 946 |
+
"loss": 0.477,
|
| 947 |
+
"step": 134
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 1.0,
|
| 951 |
+
"grad_norm": 0.6766338838361023,
|
| 952 |
+
"learning_rate": 4.077015604633669e-05,
|
| 953 |
+
"loss": 0.4422,
|
| 954 |
+
"step": 135
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 1.0074534161490682,
|
| 958 |
+
"grad_norm": 1.3875151975518343,
|
| 959 |
+
"learning_rate": 4.0619882241218684e-05,
|
| 960 |
+
"loss": 0.3755,
|
| 961 |
+
"step": 136
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 1.0149068322981367,
|
| 965 |
+
"grad_norm": 0.7353931983255126,
|
| 966 |
+
"learning_rate": 4.046871641059903e-05,
|
| 967 |
+
"loss": 0.3713,
|
| 968 |
+
"step": 137
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 1.022360248447205,
|
| 972 |
+
"grad_norm": 0.7602599418526752,
|
| 973 |
+
"learning_rate": 4.031666883229678e-05,
|
| 974 |
+
"loss": 0.3719,
|
| 975 |
+
"step": 138
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 1.0298136645962732,
|
| 979 |
+
"grad_norm": 0.5465117063921796,
|
| 980 |
+
"learning_rate": 4.016374984408137e-05,
|
| 981 |
+
"loss": 0.3602,
|
| 982 |
+
"step": 139
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 1.0372670807453417,
|
| 986 |
+
"grad_norm": 0.5515849784244337,
|
| 987 |
+
"learning_rate": 4.000996984296967e-05,
|
| 988 |
+
"loss": 0.3669,
|
| 989 |
+
"step": 140
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 1.04472049689441,
|
| 993 |
+
"grad_norm": 0.547418960473341,
|
| 994 |
+
"learning_rate": 3.985533928451914e-05,
|
| 995 |
+
"loss": 0.3565,
|
| 996 |
+
"step": 141
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 1.0521739130434782,
|
| 1000 |
+
"grad_norm": 0.43709687343518044,
|
| 1001 |
+
"learning_rate": 3.969986868211693e-05,
|
| 1002 |
+
"loss": 0.3592,
|
| 1003 |
+
"step": 142
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 1.0596273291925467,
|
| 1007 |
+
"grad_norm": 0.510533171129714,
|
| 1008 |
+
"learning_rate": 3.954356860626509e-05,
|
| 1009 |
+
"loss": 0.3548,
|
| 1010 |
+
"step": 143
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 1.067080745341615,
|
| 1014 |
+
"grad_norm": 0.7014866699955652,
|
| 1015 |
+
"learning_rate": 3.938644968386188e-05,
|
| 1016 |
+
"loss": 0.3458,
|
| 1017 |
+
"step": 144
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 1.0745341614906831,
|
| 1021 |
+
"grad_norm": 0.6054199711863072,
|
| 1022 |
+
"learning_rate": 3.922852259747921e-05,
|
| 1023 |
+
"loss": 0.3717,
|
| 1024 |
+
"step": 145
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 1.0819875776397516,
|
| 1028 |
+
"grad_norm": 0.5885498856638683,
|
| 1029 |
+
"learning_rate": 3.9069798084636357e-05,
|
| 1030 |
+
"loss": 0.3563,
|
| 1031 |
+
"step": 146
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 1.0894409937888199,
|
| 1035 |
+
"grad_norm": 0.6718059622184547,
|
| 1036 |
+
"learning_rate": 3.8910286937069894e-05,
|
| 1037 |
+
"loss": 0.3396,
|
| 1038 |
+
"step": 147
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 1.0968944099378881,
|
| 1042 |
+
"grad_norm": 0.6261142587490304,
|
| 1043 |
+
"learning_rate": 3.875e-05,
|
| 1044 |
+
"loss": 0.3675,
|
| 1045 |
+
"step": 148
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 1.1043478260869566,
|
| 1049 |
+
"grad_norm": 0.7988169272624631,
|
| 1050 |
+
"learning_rate": 3.858894817139304e-05,
|
| 1051 |
+
"loss": 0.3475,
|
| 1052 |
+
"step": 149
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 1.1118012422360248,
|
| 1056 |
+
"grad_norm": 0.5931174308761531,
|
| 1057 |
+
"learning_rate": 3.8427142401220634e-05,
|
| 1058 |
+
"loss": 0.3475,
|
| 1059 |
+
"step": 150
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 1.119254658385093,
|
| 1063 |
+
"grad_norm": 0.6442383924397598,
|
| 1064 |
+
"learning_rate": 3.8264593690715165e-05,
|
| 1065 |
+
"loss": 0.3497,
|
| 1066 |
+
"step": 151
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 1.1267080745341616,
|
| 1070 |
+
"grad_norm": 0.5725732820991489,
|
| 1071 |
+
"learning_rate": 3.810131309162178e-05,
|
| 1072 |
+
"loss": 0.3559,
|
| 1073 |
+
"step": 152
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 1.1341614906832298,
|
| 1077 |
+
"grad_norm": 0.5564507237104445,
|
| 1078 |
+
"learning_rate": 3.7937311705447016e-05,
|
| 1079 |
+
"loss": 0.3626,
|
| 1080 |
+
"step": 153
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 1.141614906832298,
|
| 1084 |
+
"grad_norm": 0.6381547420777212,
|
| 1085 |
+
"learning_rate": 3.777260068270396e-05,
|
| 1086 |
+
"loss": 0.3405,
|
| 1087 |
+
"step": 154
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 1.1490683229813665,
|
| 1091 |
+
"grad_norm": 0.5057416039471847,
|
| 1092 |
+
"learning_rate": 3.760719122215416e-05,
|
| 1093 |
+
"loss": 0.3524,
|
| 1094 |
+
"step": 155
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 1.1565217391304348,
|
| 1098 |
+
"grad_norm": 0.5412142450588553,
|
| 1099 |
+
"learning_rate": 3.74410945700462e-05,
|
| 1100 |
+
"loss": 0.3513,
|
| 1101 |
+
"step": 156
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 1.163975155279503,
|
| 1105 |
+
"grad_norm": 0.5764227357077137,
|
| 1106 |
+
"learning_rate": 3.727432201935107e-05,
|
| 1107 |
+
"loss": 0.349,
|
| 1108 |
+
"step": 157
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 1.1714285714285715,
|
| 1112 |
+
"grad_norm": 0.43438798608153883,
|
| 1113 |
+
"learning_rate": 3.7106884908994314e-05,
|
| 1114 |
+
"loss": 0.3526,
|
| 1115 |
+
"step": 158
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 1.1788819875776397,
|
| 1119 |
+
"grad_norm": 0.4757449865488938,
|
| 1120 |
+
"learning_rate": 3.693879462308516e-05,
|
| 1121 |
+
"loss": 0.3554,
|
| 1122 |
+
"step": 159
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 1.186335403726708,
|
| 1126 |
+
"grad_norm": 0.5422686029816232,
|
| 1127 |
+
"learning_rate": 3.677006259014247e-05,
|
| 1128 |
+
"loss": 0.3449,
|
| 1129 |
+
"step": 160
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 1.1937888198757765,
|
| 1133 |
+
"grad_norm": 0.4828269978461137,
|
| 1134 |
+
"learning_rate": 3.6600700282317704e-05,
|
| 1135 |
+
"loss": 0.3442,
|
| 1136 |
+
"step": 161
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 1.2012422360248447,
|
| 1140 |
+
"grad_norm": 0.6064830533601335,
|
| 1141 |
+
"learning_rate": 3.643071921461497e-05,
|
| 1142 |
+
"loss": 0.3602,
|
| 1143 |
+
"step": 162
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 1.208695652173913,
|
| 1147 |
+
"grad_norm": 0.37720957669313665,
|
| 1148 |
+
"learning_rate": 3.626013094410803e-05,
|
| 1149 |
+
"loss": 0.3632,
|
| 1150 |
+
"step": 163
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 1.2161490683229814,
|
| 1154 |
+
"grad_norm": 0.5304204188881894,
|
| 1155 |
+
"learning_rate": 3.6088947069154624e-05,
|
| 1156 |
+
"loss": 0.3555,
|
| 1157 |
+
"step": 164
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 1.2236024844720497,
|
| 1161 |
+
"grad_norm": 0.38241688724581824,
|
| 1162 |
+
"learning_rate": 3.591717922860785e-05,
|
| 1163 |
+
"loss": 0.3461,
|
| 1164 |
+
"step": 165
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 1.231055900621118,
|
| 1168 |
+
"grad_norm": 0.4132634739723132,
|
| 1169 |
+
"learning_rate": 3.574483910102481e-05,
|
| 1170 |
+
"loss": 0.3473,
|
| 1171 |
+
"step": 166
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 1.2385093167701864,
|
| 1175 |
+
"grad_norm": 0.41939353269128443,
|
| 1176 |
+
"learning_rate": 3.557193840387263e-05,
|
| 1177 |
+
"loss": 0.3592,
|
| 1178 |
+
"step": 167
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 1.2459627329192546,
|
| 1182 |
+
"grad_norm": 0.4076888356556453,
|
| 1183 |
+
"learning_rate": 3.539848889273175e-05,
|
| 1184 |
+
"loss": 0.3503,
|
| 1185 |
+
"step": 168
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 1.253416149068323,
|
| 1189 |
+
"grad_norm": 0.4270944790896845,
|
| 1190 |
+
"learning_rate": 3.522450236049668e-05,
|
| 1191 |
+
"loss": 0.3541,
|
| 1192 |
+
"step": 169
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 1.2608695652173914,
|
| 1196 |
+
"grad_norm": 0.463757861659065,
|
| 1197 |
+
"learning_rate": 3.504999063657418e-05,
|
| 1198 |
+
"loss": 0.3542,
|
| 1199 |
+
"step": 170
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 1.2683229813664596,
|
| 1203 |
+
"grad_norm": 0.3981674010803556,
|
| 1204 |
+
"learning_rate": 3.487496558607898e-05,
|
| 1205 |
+
"loss": 0.3487,
|
| 1206 |
+
"step": 171
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 1.275776397515528,
|
| 1210 |
+
"grad_norm": 0.41590787708330634,
|
| 1211 |
+
"learning_rate": 3.4699439109027074e-05,
|
| 1212 |
+
"loss": 0.3498,
|
| 1213 |
+
"step": 172
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 1.2832298136645963,
|
| 1217 |
+
"grad_norm": 0.44072247076963306,
|
| 1218 |
+
"learning_rate": 3.452342313952662e-05,
|
| 1219 |
+
"loss": 0.3525,
|
| 1220 |
+
"step": 173
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 1.2906832298136646,
|
| 1224 |
+
"grad_norm": 0.38745116662865536,
|
| 1225 |
+
"learning_rate": 3.4346929644966564e-05,
|
| 1226 |
+
"loss": 0.3546,
|
| 1227 |
+
"step": 174
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 1.298136645962733,
|
| 1231 |
+
"grad_norm": 0.40585109654853857,
|
| 1232 |
+
"learning_rate": 3.416997062520292e-05,
|
| 1233 |
+
"loss": 0.3508,
|
| 1234 |
+
"step": 175
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 1.3055900621118013,
|
| 1238 |
+
"grad_norm": 0.4410011484738261,
|
| 1239 |
+
"learning_rate": 3.399255811174295e-05,
|
| 1240 |
+
"loss": 0.351,
|
| 1241 |
+
"step": 176
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 1.3130434782608695,
|
| 1245 |
+
"grad_norm": 0.4196810442218097,
|
| 1246 |
+
"learning_rate": 3.38147041669271e-05,
|
| 1247 |
+
"loss": 0.3678,
|
| 1248 |
+
"step": 177
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 1.320496894409938,
|
| 1252 |
+
"grad_norm": 0.3579720168253587,
|
| 1253 |
+
"learning_rate": 3.36364208831089e-05,
|
| 1254 |
+
"loss": 0.3585,
|
| 1255 |
+
"step": 178
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 1.3279503105590063,
|
| 1259 |
+
"grad_norm": 0.428787604095252,
|
| 1260 |
+
"learning_rate": 3.345772038183281e-05,
|
| 1261 |
+
"loss": 0.3481,
|
| 1262 |
+
"step": 179
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 1.3354037267080745,
|
| 1266 |
+
"grad_norm": 0.3731847145526466,
|
| 1267 |
+
"learning_rate": 3.3278614813010034e-05,
|
| 1268 |
+
"loss": 0.3529,
|
| 1269 |
+
"step": 180
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 1.342857142857143,
|
| 1273 |
+
"grad_norm": 0.4426288735756756,
|
| 1274 |
+
"learning_rate": 3.309911635409246e-05,
|
| 1275 |
+
"loss": 0.3661,
|
| 1276 |
+
"step": 181
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 1.3503105590062112,
|
| 1280 |
+
"grad_norm": 0.3937030484260621,
|
| 1281 |
+
"learning_rate": 3.291923720924473e-05,
|
| 1282 |
+
"loss": 0.3501,
|
| 1283 |
+
"step": 182
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 1.3577639751552795,
|
| 1287 |
+
"grad_norm": 0.40353386330372387,
|
| 1288 |
+
"learning_rate": 3.273898960851443e-05,
|
| 1289 |
+
"loss": 0.3537,
|
| 1290 |
+
"step": 183
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 1.365217391304348,
|
| 1294 |
+
"grad_norm": 0.3721891152141779,
|
| 1295 |
+
"learning_rate": 3.2558385807000654e-05,
|
| 1296 |
+
"loss": 0.3596,
|
| 1297 |
+
"step": 184
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 1.3726708074534162,
|
| 1301 |
+
"grad_norm": 0.36845786723670104,
|
| 1302 |
+
"learning_rate": 3.2377438084020644e-05,
|
| 1303 |
+
"loss": 0.3532,
|
| 1304 |
+
"step": 185
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 1.3801242236024844,
|
| 1308 |
+
"grad_norm": 0.43575242497786837,
|
| 1309 |
+
"learning_rate": 3.219615874227504e-05,
|
| 1310 |
+
"loss": 0.3581,
|
| 1311 |
+
"step": 186
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 1.387577639751553,
|
| 1315 |
+
"grad_norm": 0.3972116108444874,
|
| 1316 |
+
"learning_rate": 3.2014560107011324e-05,
|
| 1317 |
+
"loss": 0.3544,
|
| 1318 |
+
"step": 187
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 1.3950310559006212,
|
| 1322 |
+
"grad_norm": 0.3668661244872624,
|
| 1323 |
+
"learning_rate": 3.18326545251859e-05,
|
| 1324 |
+
"loss": 0.3551,
|
| 1325 |
+
"step": 188
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 1.4024844720496894,
|
| 1329 |
+
"grad_norm": 0.35145343781990773,
|
| 1330 |
+
"learning_rate": 3.165045436462451e-05,
|
| 1331 |
+
"loss": 0.3502,
|
| 1332 |
+
"step": 189
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 1.4099378881987579,
|
| 1336 |
+
"grad_norm": 0.3698602702933315,
|
| 1337 |
+
"learning_rate": 3.146797201318149e-05,
|
| 1338 |
+
"loss": 0.3473,
|
| 1339 |
+
"step": 190
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 1.4173913043478261,
|
| 1343 |
+
"grad_norm": 0.40228974580805166,
|
| 1344 |
+
"learning_rate": 3.128521987789738e-05,
|
| 1345 |
+
"loss": 0.36,
|
| 1346 |
+
"step": 191
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 1.4248447204968944,
|
| 1350 |
+
"grad_norm": 0.36657081099285116,
|
| 1351 |
+
"learning_rate": 3.110221038415545e-05,
|
| 1352 |
+
"loss": 0.3364,
|
| 1353 |
+
"step": 192
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 1.4322981366459628,
|
| 1357 |
+
"grad_norm": 0.4544371717792479,
|
| 1358 |
+
"learning_rate": 3.0918955974836846e-05,
|
| 1359 |
+
"loss": 0.3521,
|
| 1360 |
+
"step": 193
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 1.439751552795031,
|
| 1364 |
+
"grad_norm": 0.41609260719942065,
|
| 1365 |
+
"learning_rate": 3.073546910947461e-05,
|
| 1366 |
+
"loss": 0.3447,
|
| 1367 |
+
"step": 194
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 1.4472049689440993,
|
| 1371 |
+
"grad_norm": 0.42225853754054415,
|
| 1372 |
+
"learning_rate": 3.0551762263406576e-05,
|
| 1373 |
+
"loss": 0.3579,
|
| 1374 |
+
"step": 195
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 1.4546583850931678,
|
| 1378 |
+
"grad_norm": 0.4344358503994225,
|
| 1379 |
+
"learning_rate": 3.0367847926927134e-05,
|
| 1380 |
+
"loss": 0.3537,
|
| 1381 |
+
"step": 196
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 1.462111801242236,
|
| 1385 |
+
"grad_norm": 0.4160571014028236,
|
| 1386 |
+
"learning_rate": 3.0183738604438006e-05,
|
| 1387 |
+
"loss": 0.3529,
|
| 1388 |
+
"step": 197
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 1.4695652173913043,
|
| 1392 |
+
"grad_norm": 0.38345984224367013,
|
| 1393 |
+
"learning_rate": 2.999944681359811e-05,
|
| 1394 |
+
"loss": 0.3689,
|
| 1395 |
+
"step": 198
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 1.4770186335403728,
|
| 1399 |
+
"grad_norm": 0.40710896458852786,
|
| 1400 |
+
"learning_rate": 2.9814985084472418e-05,
|
| 1401 |
+
"loss": 0.3596,
|
| 1402 |
+
"step": 199
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 1.484472049689441,
|
| 1406 |
+
"grad_norm": 0.36671518555061516,
|
| 1407 |
+
"learning_rate": 2.9630365958680107e-05,
|
| 1408 |
+
"loss": 0.3622,
|
| 1409 |
+
"step": 200
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 1.4919254658385093,
|
| 1413 |
+
"grad_norm": 0.4044241295664962,
|
| 1414 |
+
"learning_rate": 2.9445601988541782e-05,
|
| 1415 |
+
"loss": 0.3423,
|
| 1416 |
+
"step": 201
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 1.4993788819875777,
|
| 1420 |
+
"grad_norm": 0.4084788856944912,
|
| 1421 |
+
"learning_rate": 2.9260705736226075e-05,
|
| 1422 |
+
"loss": 0.3523,
|
| 1423 |
+
"step": 202
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 1.506832298136646,
|
| 1427 |
+
"grad_norm": 0.37362508421237023,
|
| 1428 |
+
"learning_rate": 2.9075689772895538e-05,
|
| 1429 |
+
"loss": 0.342,
|
| 1430 |
+
"step": 203
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 1.5142857142857142,
|
| 1434 |
+
"grad_norm": 0.3760546446289586,
|
| 1435 |
+
"learning_rate": 2.88905666778519e-05,
|
| 1436 |
+
"loss": 0.3472,
|
| 1437 |
+
"step": 204
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 1.5217391304347827,
|
| 1441 |
+
"grad_norm": 0.34422522904634617,
|
| 1442 |
+
"learning_rate": 2.870534903768082e-05,
|
| 1443 |
+
"loss": 0.3435,
|
| 1444 |
+
"step": 205
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 1.529192546583851,
|
| 1448 |
+
"grad_norm": 0.36637523532571264,
|
| 1449 |
+
"learning_rate": 2.852004944539614e-05,
|
| 1450 |
+
"loss": 0.3576,
|
| 1451 |
+
"step": 206
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 1.5366459627329192,
|
| 1455 |
+
"grad_norm": 0.4326264584972339,
|
| 1456 |
+
"learning_rate": 2.8334680499583617e-05,
|
| 1457 |
+
"loss": 0.3509,
|
| 1458 |
+
"step": 207
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 1.5440993788819877,
|
| 1462 |
+
"grad_norm": 0.36854187428297824,
|
| 1463 |
+
"learning_rate": 2.814925480354441e-05,
|
| 1464 |
+
"loss": 0.3598,
|
| 1465 |
+
"step": 208
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 1.551552795031056,
|
| 1469 |
+
"grad_norm": 0.43487065325981833,
|
| 1470 |
+
"learning_rate": 2.7963784964438122e-05,
|
| 1471 |
+
"loss": 0.3453,
|
| 1472 |
+
"step": 209
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 1.5590062111801242,
|
| 1476 |
+
"grad_norm": 0.37541698341479846,
|
| 1477 |
+
"learning_rate": 2.777828359242567e-05,
|
| 1478 |
+
"loss": 0.3498,
|
| 1479 |
+
"step": 210
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 1.5664596273291926,
|
| 1483 |
+
"grad_norm": 0.4124753900399903,
|
| 1484 |
+
"learning_rate": 2.759276329981191e-05,
|
| 1485 |
+
"loss": 0.3532,
|
| 1486 |
+
"step": 211
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 1.5739130434782609,
|
| 1490 |
+
"grad_norm": 0.37293434521740576,
|
| 1491 |
+
"learning_rate": 2.74072367001881e-05,
|
| 1492 |
+
"loss": 0.3485,
|
| 1493 |
+
"step": 212
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 1.5813664596273291,
|
| 1497 |
+
"grad_norm": 0.4288037014823596,
|
| 1498 |
+
"learning_rate": 2.722171640757434e-05,
|
| 1499 |
+
"loss": 0.3532,
|
| 1500 |
+
"step": 213
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 1.5888198757763976,
|
| 1504 |
+
"grad_norm": 0.36312626249643093,
|
| 1505 |
+
"learning_rate": 2.703621503556189e-05,
|
| 1506 |
+
"loss": 0.3652,
|
| 1507 |
+
"step": 214
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 1.5962732919254659,
|
| 1511 |
+
"grad_norm": 0.47318947905865477,
|
| 1512 |
+
"learning_rate": 2.6850745196455594e-05,
|
| 1513 |
+
"loss": 0.3546,
|
| 1514 |
+
"step": 215
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 1.603726708074534,
|
| 1518 |
+
"grad_norm": 0.3525081596484065,
|
| 1519 |
+
"learning_rate": 2.6665319500416385e-05,
|
| 1520 |
+
"loss": 0.3503,
|
| 1521 |
+
"step": 216
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 1.6111801242236026,
|
| 1525 |
+
"grad_norm": 0.45691496651133917,
|
| 1526 |
+
"learning_rate": 2.6479950554603862e-05,
|
| 1527 |
+
"loss": 0.3483,
|
| 1528 |
+
"step": 217
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"epoch": 1.6186335403726708,
|
| 1532 |
+
"grad_norm": 0.31443317345218114,
|
| 1533 |
+
"learning_rate": 2.6294650962319177e-05,
|
| 1534 |
+
"loss": 0.3639,
|
| 1535 |
+
"step": 218
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 1.626086956521739,
|
| 1539 |
+
"grad_norm": 0.3939981518889869,
|
| 1540 |
+
"learning_rate": 2.6109433322148112e-05,
|
| 1541 |
+
"loss": 0.3591,
|
| 1542 |
+
"step": 219
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 1.6335403726708075,
|
| 1546 |
+
"grad_norm": 0.34489071813729033,
|
| 1547 |
+
"learning_rate": 2.592431022710447e-05,
|
| 1548 |
+
"loss": 0.3524,
|
| 1549 |
+
"step": 220
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 1.6409937888198758,
|
| 1553 |
+
"grad_norm": 0.35520259178429653,
|
| 1554 |
+
"learning_rate": 2.5739294263773933e-05,
|
| 1555 |
+
"loss": 0.3516,
|
| 1556 |
+
"step": 221
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 1.648447204968944,
|
| 1560 |
+
"grad_norm": 0.3388748252085296,
|
| 1561 |
+
"learning_rate": 2.555439801145823e-05,
|
| 1562 |
+
"loss": 0.3586,
|
| 1563 |
+
"step": 222
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 1.6559006211180125,
|
| 1567 |
+
"grad_norm": 0.34569372832787504,
|
| 1568 |
+
"learning_rate": 2.5369634041319895e-05,
|
| 1569 |
+
"loss": 0.353,
|
| 1570 |
+
"step": 223
|
| 1571 |
+
},
|
| 1572 |
+
{
|
| 1573 |
+
"epoch": 1.6633540372670808,
|
| 1574 |
+
"grad_norm": 0.3945538297720147,
|
| 1575 |
+
"learning_rate": 2.5185014915527587e-05,
|
| 1576 |
+
"loss": 0.3527,
|
| 1577 |
+
"step": 224
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 1.670807453416149,
|
| 1581 |
+
"grad_norm": 0.3762612151126058,
|
| 1582 |
+
"learning_rate": 2.50005531864019e-05,
|
| 1583 |
+
"loss": 0.3489,
|
| 1584 |
+
"step": 225
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 1.6782608695652175,
|
| 1588 |
+
"grad_norm": 0.44476348389377207,
|
| 1589 |
+
"learning_rate": 2.4816261395562003e-05,
|
| 1590 |
+
"loss": 0.3563,
|
| 1591 |
+
"step": 226
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 1.6857142857142857,
|
| 1595 |
+
"grad_norm": 0.3646157307605566,
|
| 1596 |
+
"learning_rate": 2.4632152073072878e-05,
|
| 1597 |
+
"loss": 0.3532,
|
| 1598 |
+
"step": 227
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 1.693167701863354,
|
| 1602 |
+
"grad_norm": 0.39063038331746,
|
| 1603 |
+
"learning_rate": 2.4448237736593422e-05,
|
| 1604 |
+
"loss": 0.3518,
|
| 1605 |
+
"step": 228
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 1.7006211180124224,
|
| 1609 |
+
"grad_norm": 0.36030888961041857,
|
| 1610 |
+
"learning_rate": 2.4264530890525395e-05,
|
| 1611 |
+
"loss": 0.3514,
|
| 1612 |
+
"step": 229
|
| 1613 |
+
},
|
| 1614 |
+
{
|
| 1615 |
+
"epoch": 1.7080745341614907,
|
| 1616 |
+
"grad_norm": 0.34694110159996094,
|
| 1617 |
+
"learning_rate": 2.408104402516317e-05,
|
| 1618 |
+
"loss": 0.3537,
|
| 1619 |
+
"step": 230
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 1.715527950310559,
|
| 1623 |
+
"grad_norm": 0.35170013263514505,
|
| 1624 |
+
"learning_rate": 2.3897789615844557e-05,
|
| 1625 |
+
"loss": 0.3556,
|
| 1626 |
+
"step": 231
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 1.7229813664596274,
|
| 1630 |
+
"grad_norm": 0.339176023939733,
|
| 1631 |
+
"learning_rate": 2.3714780122102626e-05,
|
| 1632 |
+
"loss": 0.342,
|
| 1633 |
+
"step": 232
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 1.7304347826086957,
|
| 1637 |
+
"grad_norm": 0.35440772273510307,
|
| 1638 |
+
"learning_rate": 2.3532027986818518e-05,
|
| 1639 |
+
"loss": 0.3431,
|
| 1640 |
+
"step": 233
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 1.737888198757764,
|
| 1644 |
+
"grad_norm": 0.3624908891100353,
|
| 1645 |
+
"learning_rate": 2.3349545635375498e-05,
|
| 1646 |
+
"loss": 0.3549,
|
| 1647 |
+
"step": 234
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 1.7453416149068324,
|
| 1651 |
+
"grad_norm": 0.3198798531907377,
|
| 1652 |
+
"learning_rate": 2.3167345474814118e-05,
|
| 1653 |
+
"loss": 0.358,
|
| 1654 |
+
"step": 235
|
| 1655 |
+
},
|
| 1656 |
+
{
|
| 1657 |
+
"epoch": 1.7527950310559006,
|
| 1658 |
+
"grad_norm": 0.33748379039276893,
|
| 1659 |
+
"learning_rate": 2.298543989298867e-05,
|
| 1660 |
+
"loss": 0.352,
|
| 1661 |
+
"step": 236
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 1.7602484472049689,
|
| 1665 |
+
"grad_norm": 0.35787480655458953,
|
| 1666 |
+
"learning_rate": 2.2803841257724962e-05,
|
| 1667 |
+
"loss": 0.3517,
|
| 1668 |
+
"step": 237
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 1.7677018633540373,
|
| 1672 |
+
"grad_norm": 0.3263237153649648,
|
| 1673 |
+
"learning_rate": 2.2622561915979358e-05,
|
| 1674 |
+
"loss": 0.3554,
|
| 1675 |
+
"step": 238
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 1.7751552795031056,
|
| 1679 |
+
"grad_norm": 0.381086077893375,
|
| 1680 |
+
"learning_rate": 2.2441614192999354e-05,
|
| 1681 |
+
"loss": 0.3614,
|
| 1682 |
+
"step": 239
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 1.7826086956521738,
|
| 1686 |
+
"grad_norm": 0.3615867584131201,
|
| 1687 |
+
"learning_rate": 2.226101039148557e-05,
|
| 1688 |
+
"loss": 0.3521,
|
| 1689 |
+
"step": 240
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 1.7900621118012423,
|
| 1693 |
+
"grad_norm": 0.33249983999528465,
|
| 1694 |
+
"learning_rate": 2.2080762790755282e-05,
|
| 1695 |
+
"loss": 0.35,
|
| 1696 |
+
"step": 241
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 1.7975155279503106,
|
| 1700 |
+
"grad_norm": 0.3466313876235434,
|
| 1701 |
+
"learning_rate": 2.1900883645907545e-05,
|
| 1702 |
+
"loss": 0.3478,
|
| 1703 |
+
"step": 242
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 1.8049689440993788,
|
| 1707 |
+
"grad_norm": 0.3545791130868414,
|
| 1708 |
+
"learning_rate": 2.1721385186989978e-05,
|
| 1709 |
+
"loss": 0.3383,
|
| 1710 |
+
"step": 243
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 1.8124223602484473,
|
| 1714 |
+
"grad_norm": 0.3433493456392416,
|
| 1715 |
+
"learning_rate": 2.1542279618167194e-05,
|
| 1716 |
+
"loss": 0.357,
|
| 1717 |
+
"step": 244
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 1.8198757763975155,
|
| 1721 |
+
"grad_norm": 0.35215043297702653,
|
| 1722 |
+
"learning_rate": 2.13635791168911e-05,
|
| 1723 |
+
"loss": 0.3584,
|
| 1724 |
+
"step": 245
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 1.8273291925465838,
|
| 1728 |
+
"grad_norm": 0.3306587366051955,
|
| 1729 |
+
"learning_rate": 2.1185295833072914e-05,
|
| 1730 |
+
"loss": 0.3536,
|
| 1731 |
+
"step": 246
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 1.8347826086956522,
|
| 1735 |
+
"grad_norm": 0.37995802267618,
|
| 1736 |
+
"learning_rate": 2.1007441888257055e-05,
|
| 1737 |
+
"loss": 0.3555,
|
| 1738 |
+
"step": 247
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 1.8422360248447205,
|
| 1742 |
+
"grad_norm": 0.3211766118761742,
|
| 1743 |
+
"learning_rate": 2.0830029374797085e-05,
|
| 1744 |
+
"loss": 0.3302,
|
| 1745 |
+
"step": 248
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 1.8496894409937887,
|
| 1749 |
+
"grad_norm": 0.34496506016293466,
|
| 1750 |
+
"learning_rate": 2.0653070355033438e-05,
|
| 1751 |
+
"loss": 0.3331,
|
| 1752 |
+
"step": 249
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 1.8571428571428572,
|
| 1756 |
+
"grad_norm": 0.35158994392589726,
|
| 1757 |
+
"learning_rate": 2.047657686047338e-05,
|
| 1758 |
+
"loss": 0.3522,
|
| 1759 |
+
"step": 250
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 1.8645962732919255,
|
| 1763 |
+
"grad_norm": 0.2997777930152674,
|
| 1764 |
+
"learning_rate": 2.030056089097293e-05,
|
| 1765 |
+
"loss": 0.345,
|
| 1766 |
+
"step": 251
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 1.8720496894409937,
|
| 1770 |
+
"grad_norm": 0.36933604277512955,
|
| 1771 |
+
"learning_rate": 2.0125034413921024e-05,
|
| 1772 |
+
"loss": 0.3553,
|
| 1773 |
+
"step": 252
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 1.8795031055900622,
|
| 1777 |
+
"grad_norm": 0.30278033326279763,
|
| 1778 |
+
"learning_rate": 1.9950009363425827e-05,
|
| 1779 |
+
"loss": 0.3504,
|
| 1780 |
+
"step": 253
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 1.8869565217391304,
|
| 1784 |
+
"grad_norm": 0.3317258855093456,
|
| 1785 |
+
"learning_rate": 1.9775497639503325e-05,
|
| 1786 |
+
"loss": 0.3637,
|
| 1787 |
+
"step": 254
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 1.8944099378881987,
|
| 1791 |
+
"grad_norm": 0.32228882411814175,
|
| 1792 |
+
"learning_rate": 1.9601511107268255e-05,
|
| 1793 |
+
"loss": 0.3553,
|
| 1794 |
+
"step": 255
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 1.9018633540372671,
|
| 1798 |
+
"grad_norm": 0.31378115419623065,
|
| 1799 |
+
"learning_rate": 1.9428061596127383e-05,
|
| 1800 |
+
"loss": 0.347,
|
| 1801 |
+
"step": 256
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 1.9093167701863354,
|
| 1805 |
+
"grad_norm": 0.32150232133903023,
|
| 1806 |
+
"learning_rate": 1.9255160898975195e-05,
|
| 1807 |
+
"loss": 0.3416,
|
| 1808 |
+
"step": 257
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 1.9167701863354036,
|
| 1812 |
+
"grad_norm": 0.30317071268510376,
|
| 1813 |
+
"learning_rate": 1.9082820771392157e-05,
|
| 1814 |
+
"loss": 0.3488,
|
| 1815 |
+
"step": 258
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 1.924223602484472,
|
| 1819 |
+
"grad_norm": 0.3160493816316982,
|
| 1820 |
+
"learning_rate": 1.891105293084538e-05,
|
| 1821 |
+
"loss": 0.3524,
|
| 1822 |
+
"step": 259
|
| 1823 |
+
},
|
| 1824 |
+
{
|
| 1825 |
+
"epoch": 1.9316770186335404,
|
| 1826 |
+
"grad_norm": 0.31367068123245806,
|
| 1827 |
+
"learning_rate": 1.8739869055891972e-05,
|
| 1828 |
+
"loss": 0.3532,
|
| 1829 |
+
"step": 260
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 1.9391304347826086,
|
| 1833 |
+
"grad_norm": 0.32732364352375504,
|
| 1834 |
+
"learning_rate": 1.8569280785385046e-05,
|
| 1835 |
+
"loss": 0.3437,
|
| 1836 |
+
"step": 261
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 1.946583850931677,
|
| 1840 |
+
"grad_norm": 0.3044343110689431,
|
| 1841 |
+
"learning_rate": 1.8399299717682304e-05,
|
| 1842 |
+
"loss": 0.351,
|
| 1843 |
+
"step": 262
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 1.9540372670807453,
|
| 1847 |
+
"grad_norm": 0.37461713163454335,
|
| 1848 |
+
"learning_rate": 1.822993740985754e-05,
|
| 1849 |
+
"loss": 0.3491,
|
| 1850 |
+
"step": 263
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 1.9614906832298136,
|
| 1854 |
+
"grad_norm": 0.31668154420249334,
|
| 1855 |
+
"learning_rate": 1.806120537691485e-05,
|
| 1856 |
+
"loss": 0.3473,
|
| 1857 |
+
"step": 264
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 1.968944099378882,
|
| 1861 |
+
"grad_norm": 0.3012343742566849,
|
| 1862 |
+
"learning_rate": 1.7893115091005695e-05,
|
| 1863 |
+
"loss": 0.3473,
|
| 1864 |
+
"step": 265
|
| 1865 |
+
},
|
| 1866 |
+
{
|
| 1867 |
+
"epoch": 1.9763975155279503,
|
| 1868 |
+
"grad_norm": 0.39737978489918535,
|
| 1869 |
+
"learning_rate": 1.772567798064894e-05,
|
| 1870 |
+
"loss": 0.3494,
|
| 1871 |
+
"step": 266
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 1.9838509316770185,
|
| 1875 |
+
"grad_norm": 0.33805655625614117,
|
| 1876 |
+
"learning_rate": 1.7558905429953805e-05,
|
| 1877 |
+
"loss": 0.3475,
|
| 1878 |
+
"step": 267
|
| 1879 |
+
},
|
| 1880 |
+
{
|
| 1881 |
+
"epoch": 1.991304347826087,
|
| 1882 |
+
"grad_norm": 0.3322232086747,
|
| 1883 |
+
"learning_rate": 1.739280877784584e-05,
|
| 1884 |
+
"loss": 0.3308,
|
| 1885 |
+
"step": 268
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 1.9987577639751553,
|
| 1889 |
+
"grad_norm": 0.38072556291830295,
|
| 1890 |
+
"learning_rate": 1.7227399317296043e-05,
|
| 1891 |
+
"loss": 0.3425,
|
| 1892 |
+
"step": 269
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 2.0,
|
| 1896 |
+
"grad_norm": 0.38072556291830295,
|
| 1897 |
+
"learning_rate": 1.7062688294552992e-05,
|
| 1898 |
+
"loss": 0.3282,
|
| 1899 |
+
"step": 270
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 2.0074534161490685,
|
| 1903 |
+
"grad_norm": 0.7460455385201656,
|
| 1904 |
+
"learning_rate": 1.6898686908378227e-05,
|
| 1905 |
+
"loss": 0.2385,
|
| 1906 |
+
"step": 271
|
| 1907 |
+
},
|
| 1908 |
+
{
|
| 1909 |
+
"epoch": 2.0149068322981365,
|
| 1910 |
+
"grad_norm": 0.4584394574079776,
|
| 1911 |
+
"learning_rate": 1.6735406309284847e-05,
|
| 1912 |
+
"loss": 0.2325,
|
| 1913 |
+
"step": 272
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 2.022360248447205,
|
| 1917 |
+
"grad_norm": 0.35402810903371174,
|
| 1918 |
+
"learning_rate": 1.657285759877937e-05,
|
| 1919 |
+
"loss": 0.2289,
|
| 1920 |
+
"step": 273
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 2.0298136645962734,
|
| 1924 |
+
"grad_norm": 0.41508856204183064,
|
| 1925 |
+
"learning_rate": 1.6411051828606964e-05,
|
| 1926 |
+
"loss": 0.2289,
|
| 1927 |
+
"step": 274
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"epoch": 2.0372670807453415,
|
| 1931 |
+
"grad_norm": 0.5703640626631827,
|
| 1932 |
+
"learning_rate": 1.6250000000000005e-05,
|
| 1933 |
+
"loss": 0.251,
|
| 1934 |
+
"step": 275
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 2.04472049689441,
|
| 1938 |
+
"grad_norm": 0.44030030337444215,
|
| 1939 |
+
"learning_rate": 1.6089713062930108e-05,
|
| 1940 |
+
"loss": 0.2393,
|
| 1941 |
+
"step": 276
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 2.0521739130434784,
|
| 1945 |
+
"grad_norm": 0.3634716004644788,
|
| 1946 |
+
"learning_rate": 1.5930201915363652e-05,
|
| 1947 |
+
"loss": 0.2208,
|
| 1948 |
+
"step": 277
|
| 1949 |
+
},
|
| 1950 |
+
{
|
| 1951 |
+
"epoch": 2.0596273291925464,
|
| 1952 |
+
"grad_norm": 0.40263713578890487,
|
| 1953 |
+
"learning_rate": 1.5771477402520797e-05,
|
| 1954 |
+
"loss": 0.2251,
|
| 1955 |
+
"step": 278
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 2.067080745341615,
|
| 1959 |
+
"grad_norm": 0.38573706299136595,
|
| 1960 |
+
"learning_rate": 1.5613550316138116e-05,
|
| 1961 |
+
"loss": 0.2373,
|
| 1962 |
+
"step": 279
|
| 1963 |
+
},
|
| 1964 |
+
{
|
| 1965 |
+
"epoch": 2.0745341614906834,
|
| 1966 |
+
"grad_norm": 0.3886650442753063,
|
| 1967 |
+
"learning_rate": 1.5456431393734912e-05,
|
| 1968 |
+
"loss": 0.2328,
|
| 1969 |
+
"step": 280
|
| 1970 |
+
},
|
| 1971 |
+
{
|
| 1972 |
+
"epoch": 2.0819875776397514,
|
| 1973 |
+
"grad_norm": 0.3353566159721077,
|
| 1974 |
+
"learning_rate": 1.530013131788308e-05,
|
| 1975 |
+
"loss": 0.2368,
|
| 1976 |
+
"step": 281
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 2.08944099378882,
|
| 1980 |
+
"grad_norm": 0.32235589864048775,
|
| 1981 |
+
"learning_rate": 1.5144660715480877e-05,
|
| 1982 |
+
"loss": 0.2372,
|
| 1983 |
+
"step": 282
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 2.0968944099378883,
|
| 1987 |
+
"grad_norm": 0.3449572995224796,
|
| 1988 |
+
"learning_rate": 1.4990030157030343e-05,
|
| 1989 |
+
"loss": 0.2302,
|
| 1990 |
+
"step": 283
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 2.1043478260869564,
|
| 1994 |
+
"grad_norm": 0.35370627629200513,
|
| 1995 |
+
"learning_rate": 1.4836250155918632e-05,
|
| 1996 |
+
"loss": 0.2254,
|
| 1997 |
+
"step": 284
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 2.111801242236025,
|
| 2001 |
+
"grad_norm": 0.32255576283733056,
|
| 2002 |
+
"learning_rate": 1.4683331167703218e-05,
|
| 2003 |
+
"loss": 0.2304,
|
| 2004 |
+
"step": 285
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 2.1192546583850933,
|
| 2008 |
+
"grad_norm": 0.33160244908908576,
|
| 2009 |
+
"learning_rate": 1.453128358940098e-05,
|
| 2010 |
+
"loss": 0.229,
|
| 2011 |
+
"step": 286
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 2.1267080745341613,
|
| 2015 |
+
"grad_norm": 0.3342512523189852,
|
| 2016 |
+
"learning_rate": 1.4380117758781318e-05,
|
| 2017 |
+
"loss": 0.2232,
|
| 2018 |
+
"step": 287
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 2.13416149068323,
|
| 2022 |
+
"grad_norm": 0.3154712145125721,
|
| 2023 |
+
"learning_rate": 1.4229843953663313e-05,
|
| 2024 |
+
"loss": 0.2295,
|
| 2025 |
+
"step": 288
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 2.1416149068322983,
|
| 2029 |
+
"grad_norm": 0.2998522567104169,
|
| 2030 |
+
"learning_rate": 1.4080472391216898e-05,
|
| 2031 |
+
"loss": 0.2318,
|
| 2032 |
+
"step": 289
|
| 2033 |
+
},
|
| 2034 |
+
{
|
| 2035 |
+
"epoch": 2.1490683229813663,
|
| 2036 |
+
"grad_norm": 0.3120711414761453,
|
| 2037 |
+
"learning_rate": 1.3932013227268208e-05,
|
| 2038 |
+
"loss": 0.2199,
|
| 2039 |
+
"step": 290
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 2.1565217391304348,
|
| 2043 |
+
"grad_norm": 0.31545849948494736,
|
| 2044 |
+
"learning_rate": 1.3784476555609077e-05,
|
| 2045 |
+
"loss": 0.2258,
|
| 2046 |
+
"step": 291
|
| 2047 |
+
},
|
| 2048 |
+
{
|
| 2049 |
+
"epoch": 2.1639751552795032,
|
| 2050 |
+
"grad_norm": 0.31288349011845973,
|
| 2051 |
+
"learning_rate": 1.3637872407310748e-05,
|
| 2052 |
+
"loss": 0.2372,
|
| 2053 |
+
"step": 292
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"epoch": 2.1714285714285713,
|
| 2057 |
+
"grad_norm": 0.30783876364368057,
|
| 2058 |
+
"learning_rate": 1.3492210750041873e-05,
|
| 2059 |
+
"loss": 0.2273,
|
| 2060 |
+
"step": 293
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 2.1788819875776397,
|
| 2064 |
+
"grad_norm": 0.29777744790391,
|
| 2065 |
+
"learning_rate": 1.3347501487390801e-05,
|
| 2066 |
+
"loss": 0.2298,
|
| 2067 |
+
"step": 294
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 2.186335403726708,
|
| 2071 |
+
"grad_norm": 0.30475333543232624,
|
| 2072 |
+
"learning_rate": 1.3203754458192225e-05,
|
| 2073 |
+
"loss": 0.234,
|
| 2074 |
+
"step": 295
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 2.1937888198757762,
|
| 2078 |
+
"grad_norm": 0.29158802455098537,
|
| 2079 |
+
"learning_rate": 1.3060979435858259e-05,
|
| 2080 |
+
"loss": 0.2177,
|
| 2081 |
+
"step": 296
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 2.2012422360248447,
|
| 2085 |
+
"grad_norm": 0.2910747993340585,
|
| 2086 |
+
"learning_rate": 1.2919186127713885e-05,
|
| 2087 |
+
"loss": 0.2257,
|
| 2088 |
+
"step": 297
|
| 2089 |
+
},
|
| 2090 |
+
{
|
| 2091 |
+
"epoch": 2.208695652173913,
|
| 2092 |
+
"grad_norm": 0.29580086796673544,
|
| 2093 |
+
"learning_rate": 1.2778384174337025e-05,
|
| 2094 |
+
"loss": 0.2256,
|
| 2095 |
+
"step": 298
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 2.216149068322981,
|
| 2099 |
+
"grad_norm": 0.28921712308433817,
|
| 2100 |
+
"learning_rate": 1.2638583148903008e-05,
|
| 2101 |
+
"loss": 0.2321,
|
| 2102 |
+
"step": 299
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 2.2236024844720497,
|
| 2106 |
+
"grad_norm": 0.28971064806152724,
|
| 2107 |
+
"learning_rate": 1.2499792556533716e-05,
|
| 2108 |
+
"loss": 0.2319,
|
| 2109 |
+
"step": 300
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 2.231055900621118,
|
| 2113 |
+
"grad_norm": 0.2855654258873247,
|
| 2114 |
+
"learning_rate": 1.236202183365134e-05,
|
| 2115 |
+
"loss": 0.2255,
|
| 2116 |
+
"step": 301
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"epoch": 2.238509316770186,
|
| 2120 |
+
"grad_norm": 0.291778227302567,
|
| 2121 |
+
"learning_rate": 1.2225280347336763e-05,
|
| 2122 |
+
"loss": 0.2329,
|
| 2123 |
+
"step": 302
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 2.2459627329192546,
|
| 2127 |
+
"grad_norm": 0.28470732682114813,
|
| 2128 |
+
"learning_rate": 1.20895773946927e-05,
|
| 2129 |
+
"loss": 0.2263,
|
| 2130 |
+
"step": 303
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 2.253416149068323,
|
| 2134 |
+
"grad_norm": 0.27214743085323334,
|
| 2135 |
+
"learning_rate": 1.1954922202211615e-05,
|
| 2136 |
+
"loss": 0.2262,
|
| 2137 |
+
"step": 304
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 2.260869565217391,
|
| 2141 |
+
"grad_norm": 0.2783027932885126,
|
| 2142 |
+
"learning_rate": 1.1821323925148358e-05,
|
| 2143 |
+
"loss": 0.2278,
|
| 2144 |
+
"step": 305
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 2.2683229813664596,
|
| 2148 |
+
"grad_norm": 0.2940915504230925,
|
| 2149 |
+
"learning_rate": 1.1688791646897726e-05,
|
| 2150 |
+
"loss": 0.2254,
|
| 2151 |
+
"step": 306
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 2.275776397515528,
|
| 2155 |
+
"grad_norm": 0.29241392964362756,
|
| 2156 |
+
"learning_rate": 1.1557334378376882e-05,
|
| 2157 |
+
"loss": 0.2274,
|
| 2158 |
+
"step": 307
|
| 2159 |
+
},
|
| 2160 |
+
{
|
| 2161 |
+
"epoch": 2.283229813664596,
|
| 2162 |
+
"grad_norm": 0.2859155808493438,
|
| 2163 |
+
"learning_rate": 1.1426961057412672e-05,
|
| 2164 |
+
"loss": 0.2242,
|
| 2165 |
+
"step": 308
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 2.2906832298136646,
|
| 2169 |
+
"grad_norm": 0.2863824212664396,
|
| 2170 |
+
"learning_rate": 1.1297680548133993e-05,
|
| 2171 |
+
"loss": 0.2211,
|
| 2172 |
+
"step": 309
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 2.298136645962733,
|
| 2176 |
+
"grad_norm": 0.27218685640719575,
|
| 2177 |
+
"learning_rate": 1.1169501640369051e-05,
|
| 2178 |
+
"loss": 0.2166,
|
| 2179 |
+
"step": 310
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 2.305590062111801,
|
| 2183 |
+
"grad_norm": 0.2739565076313759,
|
| 2184 |
+
"learning_rate": 1.1042433049047781e-05,
|
| 2185 |
+
"loss": 0.2288,
|
| 2186 |
+
"step": 311
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 2.3130434782608695,
|
| 2190 |
+
"grad_norm": 0.2858456970822226,
|
| 2191 |
+
"learning_rate": 1.0916483413609315e-05,
|
| 2192 |
+
"loss": 0.2252,
|
| 2193 |
+
"step": 312
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 2.320496894409938,
|
| 2197 |
+
"grad_norm": 0.2730982695340446,
|
| 2198 |
+
"learning_rate": 1.079166129741455e-05,
|
| 2199 |
+
"loss": 0.2289,
|
| 2200 |
+
"step": 313
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 2.327950310559006,
|
| 2204 |
+
"grad_norm": 0.2858131409653359,
|
| 2205 |
+
"learning_rate": 1.0667975187163976e-05,
|
| 2206 |
+
"loss": 0.2308,
|
| 2207 |
+
"step": 314
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 2.3354037267080745,
|
| 2211 |
+
"grad_norm": 0.2666152133579397,
|
| 2212 |
+
"learning_rate": 1.0545433492320603e-05,
|
| 2213 |
+
"loss": 0.2278,
|
| 2214 |
+
"step": 315
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 2.342857142857143,
|
| 2218 |
+
"grad_norm": 0.2770908273244185,
|
| 2219 |
+
"learning_rate": 1.042404454453824e-05,
|
| 2220 |
+
"loss": 0.2168,
|
| 2221 |
+
"step": 316
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 2.350310559006211,
|
| 2225 |
+
"grad_norm": 0.2738838241569534,
|
| 2226 |
+
"learning_rate": 1.0303816597095004e-05,
|
| 2227 |
+
"loss": 0.2381,
|
| 2228 |
+
"step": 317
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 2.3577639751552795,
|
| 2232 |
+
"grad_norm": 0.27892490339684844,
|
| 2233 |
+
"learning_rate": 1.0184757824332187e-05,
|
| 2234 |
+
"loss": 0.2365,
|
| 2235 |
+
"step": 318
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 2.365217391304348,
|
| 2239 |
+
"grad_norm": 0.27327486293785713,
|
| 2240 |
+
"learning_rate": 1.0066876321098467e-05,
|
| 2241 |
+
"loss": 0.2262,
|
| 2242 |
+
"step": 319
|
| 2243 |
+
},
|
| 2244 |
+
{
|
| 2245 |
+
"epoch": 2.372670807453416,
|
| 2246 |
+
"grad_norm": 0.2691733028937572,
|
| 2247 |
+
"learning_rate": 9.950180102199564e-06,
|
| 2248 |
+
"loss": 0.2287,
|
| 2249 |
+
"step": 320
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 2.3801242236024844,
|
| 2253 |
+
"grad_norm": 0.2706087500883035,
|
| 2254 |
+
"learning_rate": 9.834677101853265e-06,
|
| 2255 |
+
"loss": 0.2244,
|
| 2256 |
+
"step": 321
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 2.387577639751553,
|
| 2260 |
+
"grad_norm": 0.2855306323449237,
|
| 2261 |
+
"learning_rate": 9.720375173150024e-06,
|
| 2262 |
+
"loss": 0.2349,
|
| 2263 |
+
"step": 322
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 2.395031055900621,
|
| 2267 |
+
"grad_norm": 0.26777926603611185,
|
| 2268 |
+
"learning_rate": 9.607282087518984e-06,
|
| 2269 |
+
"loss": 0.2154,
|
| 2270 |
+
"step": 323
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 2.4024844720496894,
|
| 2274 |
+
"grad_norm": 0.2821902505187217,
|
| 2275 |
+
"learning_rate": 9.495405534199617e-06,
|
| 2276 |
+
"loss": 0.2325,
|
| 2277 |
+
"step": 324
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 2.409937888198758,
|
| 2281 |
+
"grad_norm": 0.27147857463244984,
|
| 2282 |
+
"learning_rate": 9.384753119718937e-06,
|
| 2283 |
+
"loss": 0.2228,
|
| 2284 |
+
"step": 325
|
| 2285 |
+
},
|
| 2286 |
+
{
|
| 2287 |
+
"epoch": 2.417391304347826,
|
| 2288 |
+
"grad_norm": 0.2699371891385757,
|
| 2289 |
+
"learning_rate": 9.27533236737431e-06,
|
| 2290 |
+
"loss": 0.2264,
|
| 2291 |
+
"step": 326
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 2.4248447204968944,
|
| 2295 |
+
"grad_norm": 0.2762402133611527,
|
| 2296 |
+
"learning_rate": 9.167150716721954e-06,
|
| 2297 |
+
"loss": 0.2208,
|
| 2298 |
+
"step": 327
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 2.432298136645963,
|
| 2302 |
+
"grad_norm": 0.27110209952512,
|
| 2303 |
+
"learning_rate": 9.060215523071117e-06,
|
| 2304 |
+
"loss": 0.2253,
|
| 2305 |
+
"step": 328
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 2.439751552795031,
|
| 2309 |
+
"grad_norm": 0.27014197741120893,
|
| 2310 |
+
"learning_rate": 8.954534056983984e-06,
|
| 2311 |
+
"loss": 0.2293,
|
| 2312 |
+
"step": 329
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 2.4472049689440993,
|
| 2316 |
+
"grad_norm": 0.27495397443241254,
|
| 2317 |
+
"learning_rate": 8.850113503781367e-06,
|
| 2318 |
+
"loss": 0.2209,
|
| 2319 |
+
"step": 330
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 2.454658385093168,
|
| 2323 |
+
"grad_norm": 0.27468805910530714,
|
| 2324 |
+
"learning_rate": 8.746960963054145e-06,
|
| 2325 |
+
"loss": 0.2349,
|
| 2326 |
+
"step": 331
|
| 2327 |
+
},
|
| 2328 |
+
{
|
| 2329 |
+
"epoch": 2.462111801242236,
|
| 2330 |
+
"grad_norm": 0.2770381414499631,
|
| 2331 |
+
"learning_rate": 8.645083448180574e-06,
|
| 2332 |
+
"loss": 0.2238,
|
| 2333 |
+
"step": 332
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 2.4695652173913043,
|
| 2337 |
+
"grad_norm": 0.27376387230241217,
|
| 2338 |
+
"learning_rate": 8.54448788584946e-06,
|
| 2339 |
+
"loss": 0.2207,
|
| 2340 |
+
"step": 333
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 2.4770186335403728,
|
| 2344 |
+
"grad_norm": 0.27580353431299076,
|
| 2345 |
+
"learning_rate": 8.445181115589179e-06,
|
| 2346 |
+
"loss": 0.2219,
|
| 2347 |
+
"step": 334
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 2.4844720496894412,
|
| 2351 |
+
"grad_norm": 0.26703692348696817,
|
| 2352 |
+
"learning_rate": 8.34716988930267e-06,
|
| 2353 |
+
"loss": 0.2187,
|
| 2354 |
+
"step": 335
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 2.4919254658385093,
|
| 2358 |
+
"grad_norm": 0.2651902683216356,
|
| 2359 |
+
"learning_rate": 8.250460870808394e-06,
|
| 2360 |
+
"loss": 0.2238,
|
| 2361 |
+
"step": 336
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 2.4993788819875777,
|
| 2365 |
+
"grad_norm": 0.2602841629540144,
|
| 2366 |
+
"learning_rate": 8.155060635387206e-06,
|
| 2367 |
+
"loss": 0.2237,
|
| 2368 |
+
"step": 337
|
| 2369 |
+
},
|
| 2370 |
+
{
|
| 2371 |
+
"epoch": 2.506832298136646,
|
| 2372 |
+
"grad_norm": 0.26591023651045076,
|
| 2373 |
+
"learning_rate": 8.060975669335365e-06,
|
| 2374 |
+
"loss": 0.2283,
|
| 2375 |
+
"step": 338
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 2.5142857142857142,
|
| 2379 |
+
"grad_norm": 0.26252263621744065,
|
| 2380 |
+
"learning_rate": 7.968212369523462e-06,
|
| 2381 |
+
"loss": 0.2245,
|
| 2382 |
+
"step": 339
|
| 2383 |
+
},
|
| 2384 |
+
{
|
| 2385 |
+
"epoch": 2.5217391304347827,
|
| 2386 |
+
"grad_norm": 0.2748374618397896,
|
| 2387 |
+
"learning_rate": 7.876777042961544e-06,
|
| 2388 |
+
"loss": 0.2228,
|
| 2389 |
+
"step": 340
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 2.529192546583851,
|
| 2393 |
+
"grad_norm": 0.2597274246070528,
|
| 2394 |
+
"learning_rate": 7.786675906370278e-06,
|
| 2395 |
+
"loss": 0.2283,
|
| 2396 |
+
"step": 341
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 2.536645962732919,
|
| 2400 |
+
"grad_norm": 0.2687998792747167,
|
| 2401 |
+
"learning_rate": 7.697915085758266e-06,
|
| 2402 |
+
"loss": 0.2205,
|
| 2403 |
+
"step": 342
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 2.5440993788819877,
|
| 2407 |
+
"grad_norm": 0.26277761372100056,
|
| 2408 |
+
"learning_rate": 7.610500616005556e-06,
|
| 2409 |
+
"loss": 0.2201,
|
| 2410 |
+
"step": 343
|
| 2411 |
+
},
|
| 2412 |
+
{
|
| 2413 |
+
"epoch": 2.551552795031056,
|
| 2414 |
+
"grad_norm": 0.273978880458127,
|
| 2415 |
+
"learning_rate": 7.524438440453323e-06,
|
| 2416 |
+
"loss": 0.2332,
|
| 2417 |
+
"step": 344
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 2.559006211180124,
|
| 2421 |
+
"grad_norm": 0.25462514116789337,
|
| 2422 |
+
"learning_rate": 7.439734410499752e-06,
|
| 2423 |
+
"loss": 0.2188,
|
| 2424 |
+
"step": 345
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 2.5664596273291926,
|
| 2428 |
+
"grad_norm": 0.271284156037868,
|
| 2429 |
+
"learning_rate": 7.356394285202248e-06,
|
| 2430 |
+
"loss": 0.2252,
|
| 2431 |
+
"step": 346
|
| 2432 |
+
},
|
| 2433 |
+
{
|
| 2434 |
+
"epoch": 2.573913043478261,
|
| 2435 |
+
"grad_norm": 0.2709719202368051,
|
| 2436 |
+
"learning_rate": 7.274423730885835e-06,
|
| 2437 |
+
"loss": 0.2377,
|
| 2438 |
+
"step": 347
|
| 2439 |
+
},
|
| 2440 |
+
{
|
| 2441 |
+
"epoch": 2.581366459627329,
|
| 2442 |
+
"grad_norm": 0.2687629580156147,
|
| 2443 |
+
"learning_rate": 7.193828320757909e-06,
|
| 2444 |
+
"loss": 0.2212,
|
| 2445 |
+
"step": 348
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 2.5888198757763976,
|
| 2449 |
+
"grad_norm": 0.2723242179612203,
|
| 2450 |
+
"learning_rate": 7.114613534529333e-06,
|
| 2451 |
+
"loss": 0.2169,
|
| 2452 |
+
"step": 349
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 2.596273291925466,
|
| 2456 |
+
"grad_norm": 0.25830008261692217,
|
| 2457 |
+
"learning_rate": 7.036784758041846e-06,
|
| 2458 |
+
"loss": 0.2151,
|
| 2459 |
+
"step": 350
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 2.603726708074534,
|
| 2463 |
+
"grad_norm": 0.27058413732950065,
|
| 2464 |
+
"learning_rate": 6.960347282901894e-06,
|
| 2465 |
+
"loss": 0.226,
|
| 2466 |
+
"step": 351
|
| 2467 |
+
},
|
| 2468 |
+
{
|
| 2469 |
+
"epoch": 2.6111801242236026,
|
| 2470 |
+
"grad_norm": 0.2602830736394775,
|
| 2471 |
+
"learning_rate": 6.885306306120837e-06,
|
| 2472 |
+
"loss": 0.2173,
|
| 2473 |
+
"step": 352
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 2.618633540372671,
|
| 2477 |
+
"grad_norm": 0.2884065006859397,
|
| 2478 |
+
"learning_rate": 6.811666929761612e-06,
|
| 2479 |
+
"loss": 0.223,
|
| 2480 |
+
"step": 353
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"epoch": 2.626086956521739,
|
| 2484 |
+
"grad_norm": 0.2678949940900833,
|
| 2485 |
+
"learning_rate": 6.739434160591852e-06,
|
| 2486 |
+
"loss": 0.2232,
|
| 2487 |
+
"step": 354
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 2.6335403726708075,
|
| 2491 |
+
"grad_norm": 0.26613900004744706,
|
| 2492 |
+
"learning_rate": 6.668612909743448e-06,
|
| 2493 |
+
"loss": 0.2109,
|
| 2494 |
+
"step": 355
|
| 2495 |
+
},
|
| 2496 |
+
{
|
| 2497 |
+
"epoch": 2.640993788819876,
|
| 2498 |
+
"grad_norm": 0.2600134334991607,
|
| 2499 |
+
"learning_rate": 6.599207992378657e-06,
|
| 2500 |
+
"loss": 0.2216,
|
| 2501 |
+
"step": 356
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 2.648447204968944,
|
| 2505 |
+
"grad_norm": 0.26265543365401706,
|
| 2506 |
+
"learning_rate": 6.531224127362726e-06,
|
| 2507 |
+
"loss": 0.223,
|
| 2508 |
+
"step": 357
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 2.6559006211180125,
|
| 2512 |
+
"grad_norm": 0.26558869901544513,
|
| 2513 |
+
"learning_rate": 6.464665936943023e-06,
|
| 2514 |
+
"loss": 0.2172,
|
| 2515 |
+
"step": 358
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 2.663354037267081,
|
| 2519 |
+
"grad_norm": 0.2693433022214806,
|
| 2520 |
+
"learning_rate": 6.399537946434801e-06,
|
| 2521 |
+
"loss": 0.2198,
|
| 2522 |
+
"step": 359
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 2.670807453416149,
|
| 2526 |
+
"grad_norm": 0.2604150700405178,
|
| 2527 |
+
"learning_rate": 6.335844583913515e-06,
|
| 2528 |
+
"loss": 0.22,
|
| 2529 |
+
"step": 360
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 2.6782608695652175,
|
| 2533 |
+
"grad_norm": 0.26795928209419323,
|
| 2534 |
+
"learning_rate": 6.27359017991373e-06,
|
| 2535 |
+
"loss": 0.2254,
|
| 2536 |
+
"step": 361
|
| 2537 |
+
},
|
| 2538 |
+
{
|
| 2539 |
+
"epoch": 2.685714285714286,
|
| 2540 |
+
"grad_norm": 0.26825259134506735,
|
| 2541 |
+
"learning_rate": 6.212778967134715e-06,
|
| 2542 |
+
"loss": 0.2286,
|
| 2543 |
+
"step": 362
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 2.693167701863354,
|
| 2547 |
+
"grad_norm": 0.2624757714546698,
|
| 2548 |
+
"learning_rate": 6.153415080152655e-06,
|
| 2549 |
+
"loss": 0.2224,
|
| 2550 |
+
"step": 363
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 2.7006211180124224,
|
| 2554 |
+
"grad_norm": 0.2538413882658132,
|
| 2555 |
+
"learning_rate": 6.095502555139516e-06,
|
| 2556 |
+
"loss": 0.2116,
|
| 2557 |
+
"step": 364
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 2.708074534161491,
|
| 2561 |
+
"grad_norm": 0.26097610705351,
|
| 2562 |
+
"learning_rate": 6.039045329588671e-06,
|
| 2563 |
+
"loss": 0.2135,
|
| 2564 |
+
"step": 365
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 2.715527950310559,
|
| 2568 |
+
"grad_norm": 0.26841385627861275,
|
| 2569 |
+
"learning_rate": 5.984047242047134e-06,
|
| 2570 |
+
"loss": 0.2126,
|
| 2571 |
+
"step": 366
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 2.7229813664596274,
|
| 2575 |
+
"grad_norm": 0.26691120572518234,
|
| 2576 |
+
"learning_rate": 5.930512031854617e-06,
|
| 2577 |
+
"loss": 0.2231,
|
| 2578 |
+
"step": 367
|
| 2579 |
+
},
|
| 2580 |
+
{
|
| 2581 |
+
"epoch": 2.730434782608696,
|
| 2582 |
+
"grad_norm": 0.2595939680410351,
|
| 2583 |
+
"learning_rate": 5.8784433388892726e-06,
|
| 2584 |
+
"loss": 0.2187,
|
| 2585 |
+
"step": 368
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 2.737888198757764,
|
| 2589 |
+
"grad_norm": 0.2620978338502642,
|
| 2590 |
+
"learning_rate": 5.827844703320216e-06,
|
| 2591 |
+
"loss": 0.2275,
|
| 2592 |
+
"step": 369
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 2.7453416149068324,
|
| 2596 |
+
"grad_norm": 0.27293115847294985,
|
| 2597 |
+
"learning_rate": 5.778719565366846e-06,
|
| 2598 |
+
"loss": 0.2216,
|
| 2599 |
+
"step": 370
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 2.752795031055901,
|
| 2603 |
+
"grad_norm": 0.26642579246201126,
|
| 2604 |
+
"learning_rate": 5.731071265064913e-06,
|
| 2605 |
+
"loss": 0.2253,
|
| 2606 |
+
"step": 371
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 2.760248447204969,
|
| 2610 |
+
"grad_norm": 0.2673788170594958,
|
| 2611 |
+
"learning_rate": 5.684903042039452e-06,
|
| 2612 |
+
"loss": 0.2271,
|
| 2613 |
+
"step": 372
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 2.7677018633540373,
|
| 2617 |
+
"grad_norm": 0.2554915331836976,
|
| 2618 |
+
"learning_rate": 5.640218035284521e-06,
|
| 2619 |
+
"loss": 0.2295,
|
| 2620 |
+
"step": 373
|
| 2621 |
+
},
|
| 2622 |
+
{
|
| 2623 |
+
"epoch": 2.775155279503106,
|
| 2624 |
+
"grad_norm": 0.2678452830442339,
|
| 2625 |
+
"learning_rate": 5.59701928294976e-06,
|
| 2626 |
+
"loss": 0.2215,
|
| 2627 |
+
"step": 374
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 2.782608695652174,
|
| 2631 |
+
"grad_norm": 0.2815341168477551,
|
| 2632 |
+
"learning_rate": 5.555309722133842e-06,
|
| 2633 |
+
"loss": 0.2281,
|
| 2634 |
+
"step": 375
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 2.7900621118012423,
|
| 2638 |
+
"grad_norm": 0.2624788732767209,
|
| 2639 |
+
"learning_rate": 5.515092188684775e-06,
|
| 2640 |
+
"loss": 0.2247,
|
| 2641 |
+
"step": 376
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 2.7975155279503108,
|
| 2645 |
+
"grad_norm": 0.27489059149241923,
|
| 2646 |
+
"learning_rate": 5.476369417007091e-06,
|
| 2647 |
+
"loss": 0.2327,
|
| 2648 |
+
"step": 377
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 2.804968944099379,
|
| 2652 |
+
"grad_norm": 0.28772902740092976,
|
| 2653 |
+
"learning_rate": 5.439144039875931e-06,
|
| 2654 |
+
"loss": 0.2167,
|
| 2655 |
+
"step": 378
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 2.8124223602484473,
|
| 2659 |
+
"grad_norm": 0.26111982505416115,
|
| 2660 |
+
"learning_rate": 5.403418588258045e-06,
|
| 2661 |
+
"loss": 0.2197,
|
| 2662 |
+
"step": 379
|
| 2663 |
+
},
|
| 2664 |
+
{
|
| 2665 |
+
"epoch": 2.8198757763975157,
|
| 2666 |
+
"grad_norm": 0.267283694967123,
|
| 2667 |
+
"learning_rate": 5.369195491139709e-06,
|
| 2668 |
+
"loss": 0.2262,
|
| 2669 |
+
"step": 380
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 2.8273291925465838,
|
| 2673 |
+
"grad_norm": 0.27325385624236986,
|
| 2674 |
+
"learning_rate": 5.336477075361577e-06,
|
| 2675 |
+
"loss": 0.2172,
|
| 2676 |
+
"step": 381
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 2.8347826086956522,
|
| 2680 |
+
"grad_norm": 0.2647115488846027,
|
| 2681 |
+
"learning_rate": 5.305265565460477e-06,
|
| 2682 |
+
"loss": 0.2244,
|
| 2683 |
+
"step": 382
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 2.8422360248447207,
|
| 2687 |
+
"grad_norm": 0.26432835544386557,
|
| 2688 |
+
"learning_rate": 5.275563083518169e-06,
|
| 2689 |
+
"loss": 0.216,
|
| 2690 |
+
"step": 383
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 2.8496894409937887,
|
| 2694 |
+
"grad_norm": 0.2574806933503353,
|
| 2695 |
+
"learning_rate": 5.247371649017059e-06,
|
| 2696 |
+
"loss": 0.2309,
|
| 2697 |
+
"step": 384
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 2.857142857142857,
|
| 2701 |
+
"grad_norm": 0.26132718645601316,
|
| 2702 |
+
"learning_rate": 5.220693178702895e-06,
|
| 2703 |
+
"loss": 0.2142,
|
| 2704 |
+
"step": 385
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 2.8645962732919257,
|
| 2708 |
+
"grad_norm": 0.2778266622426319,
|
| 2709 |
+
"learning_rate": 5.195529486454448e-06,
|
| 2710 |
+
"loss": 0.2246,
|
| 2711 |
+
"step": 386
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 2.8720496894409937,
|
| 2715 |
+
"grad_norm": 0.2548174424842807,
|
| 2716 |
+
"learning_rate": 5.171882283160185e-06,
|
| 2717 |
+
"loss": 0.2207,
|
| 2718 |
+
"step": 387
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 2.879503105590062,
|
| 2722 |
+
"grad_norm": 0.26469737507644614,
|
| 2723 |
+
"learning_rate": 5.149753176601936e-06,
|
| 2724 |
+
"loss": 0.2245,
|
| 2725 |
+
"step": 388
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 2.8869565217391306,
|
| 2729 |
+
"grad_norm": 0.2660467641695125,
|
| 2730 |
+
"learning_rate": 5.129143671345609e-06,
|
| 2731 |
+
"loss": 0.227,
|
| 2732 |
+
"step": 389
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 2.8944099378881987,
|
| 2736 |
+
"grad_norm": 0.253693975774761,
|
| 2737 |
+
"learning_rate": 5.110055168638854e-06,
|
| 2738 |
+
"loss": 0.219,
|
| 2739 |
+
"step": 390
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 2.901863354037267,
|
| 2743 |
+
"grad_norm": 0.2795604845969123,
|
| 2744 |
+
"learning_rate": 5.09248896631582e-06,
|
| 2745 |
+
"loss": 0.2321,
|
| 2746 |
+
"step": 391
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 2.9093167701863356,
|
| 2750 |
+
"grad_norm": 0.7496235586126184,
|
| 2751 |
+
"learning_rate": 5.07644625870891e-06,
|
| 2752 |
+
"loss": 0.2366,
|
| 2753 |
+
"step": 392
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 2.9167701863354036,
|
| 2757 |
+
"grad_norm": 0.2618794070766183,
|
| 2758 |
+
"learning_rate": 5.06192813656757e-06,
|
| 2759 |
+
"loss": 0.2283,
|
| 2760 |
+
"step": 393
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 2.924223602484472,
|
| 2764 |
+
"grad_norm": 0.2616116501289447,
|
| 2765 |
+
"learning_rate": 5.048935586984133e-06,
|
| 2766 |
+
"loss": 0.2177,
|
| 2767 |
+
"step": 394
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 2.9316770186335406,
|
| 2771 |
+
"grad_norm": 0.2719537845754615,
|
| 2772 |
+
"learning_rate": 5.0374694933267114e-06,
|
| 2773 |
+
"loss": 0.2273,
|
| 2774 |
+
"step": 395
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 2.9391304347826086,
|
| 2778 |
+
"grad_norm": 0.26226288303787226,
|
| 2779 |
+
"learning_rate": 5.027530635179121e-06,
|
| 2780 |
+
"loss": 0.2132,
|
| 2781 |
+
"step": 396
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 2.946583850931677,
|
| 2785 |
+
"grad_norm": 0.24929011431595574,
|
| 2786 |
+
"learning_rate": 5.019119688287901e-06,
|
| 2787 |
+
"loss": 0.2238,
|
| 2788 |
+
"step": 397
|
| 2789 |
+
},
|
| 2790 |
+
{
|
| 2791 |
+
"epoch": 2.9540372670807455,
|
| 2792 |
+
"grad_norm": 0.2606373590411471,
|
| 2793 |
+
"learning_rate": 5.012237224516342e-06,
|
| 2794 |
+
"loss": 0.2216,
|
| 2795 |
+
"step": 398
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 2.9614906832298136,
|
| 2799 |
+
"grad_norm": 0.29892861469452736,
|
| 2800 |
+
"learning_rate": 5.00688371180563e-06,
|
| 2801 |
+
"loss": 0.2362,
|
| 2802 |
+
"step": 399
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 2.968944099378882,
|
| 2806 |
+
"grad_norm": 0.2665719018831708,
|
| 2807 |
+
"learning_rate": 5.003059514143014e-06,
|
| 2808 |
+
"loss": 0.221,
|
| 2809 |
+
"step": 400
|
| 2810 |
+
},
|
| 2811 |
+
{
|
| 2812 |
+
"epoch": 2.9763975155279505,
|
| 2813 |
+
"grad_norm": 0.25860725999612966,
|
| 2814 |
+
"learning_rate": 5.000764891537067e-06,
|
| 2815 |
+
"loss": 0.2204,
|
| 2816 |
+
"step": 401
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 2.9838509316770185,
|
| 2820 |
+
"grad_norm": 0.25161372246359515,
|
| 2821 |
+
"learning_rate": 5e-06,
|
| 2822 |
+
"loss": 0.2245,
|
| 2823 |
+
"step": 402
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 2.9838509316770185,
|
| 2827 |
+
"step": 402,
|
| 2828 |
+
"total_flos": 8.574354709534474e+17,
|
| 2829 |
+
"train_loss": 0.3579157315083404,
|
| 2830 |
+
"train_runtime": 11555.4492,
|
| 2831 |
+
"train_samples_per_second": 6.688,
|
| 2832 |
+
"train_steps_per_second": 0.035
|
| 2833 |
+
}
|
| 2834 |
+
],
|
| 2835 |
+
"logging_steps": 1,
|
| 2836 |
+
"max_steps": 402,
|
| 2837 |
+
"num_input_tokens_seen": 0,
|
| 2838 |
+
"num_train_epochs": 3,
|
| 2839 |
+
"save_steps": 100,
|
| 2840 |
+
"stateful_callbacks": {
|
| 2841 |
+
"TrainerControl": {
|
| 2842 |
+
"args": {
|
| 2843 |
+
"should_epoch_stop": false,
|
| 2844 |
+
"should_evaluate": false,
|
| 2845 |
+
"should_log": false,
|
| 2846 |
+
"should_save": true,
|
| 2847 |
+
"should_training_stop": true
|
| 2848 |
+
},
|
| 2849 |
+
"attributes": {}
|
| 2850 |
+
}
|
| 2851 |
+
},
|
| 2852 |
+
"total_flos": 8.574354709534474e+17,
|
| 2853 |
+
"train_batch_size": 4,
|
| 2854 |
+
"trial_name": null,
|
| 2855 |
+
"trial_params": null
|
| 2856 |
+
}
|