Model save
Browse files- README.md +3 -5
- all_results.json +6 -6
- train_results.json +6 -6
- trainer_state.json +1802 -45
README.md
CHANGED
|
@@ -1,11 +1,9 @@
|
|
| 1 |
---
|
| 2 |
base_model: Qwen/Qwen2.5-0.5B-Instruct
|
| 3 |
-
datasets: YangZhoumill/post_v_1
|
| 4 |
library_name: transformers
|
| 5 |
model_name: Qwen2.5-1.5B-Open-R1-Distill
|
| 6 |
tags:
|
| 7 |
- generated_from_trainer
|
| 8 |
-
- open-r1
|
| 9 |
- trl
|
| 10 |
- sft
|
| 11 |
licence: license
|
|
@@ -13,7 +11,7 @@ licence: license
|
|
| 13 |
|
| 14 |
# Model Card for Qwen2.5-1.5B-Open-R1-Distill
|
| 15 |
|
| 16 |
-
This model is a fine-tuned version of [Qwen/Qwen2.5-0.5B-Instruct](https://huggingface.co/Qwen/Qwen2.5-0.5B-Instruct)
|
| 17 |
It has been trained using [TRL](https://github.com/huggingface/trl).
|
| 18 |
|
| 19 |
## Quick start
|
|
@@ -22,14 +20,14 @@ It has been trained using [TRL](https://github.com/huggingface/trl).
|
|
| 22 |
from transformers import pipeline
|
| 23 |
|
| 24 |
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?"
|
| 25 |
-
generator = pipeline("text-generation", model="
|
| 26 |
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
|
| 27 |
print(output["generated_text"])
|
| 28 |
```
|
| 29 |
|
| 30 |
## Training procedure
|
| 31 |
|
| 32 |
-
[<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/stevenzhou0816100/huggingface/runs/
|
| 33 |
|
| 34 |
|
| 35 |
This model was trained with SFT.
|
|
|
|
| 1 |
---
|
| 2 |
base_model: Qwen/Qwen2.5-0.5B-Instruct
|
|
|
|
| 3 |
library_name: transformers
|
| 4 |
model_name: Qwen2.5-1.5B-Open-R1-Distill
|
| 5 |
tags:
|
| 6 |
- generated_from_trainer
|
|
|
|
| 7 |
- trl
|
| 8 |
- sft
|
| 9 |
licence: license
|
|
|
|
| 11 |
|
| 12 |
# Model Card for Qwen2.5-1.5B-Open-R1-Distill
|
| 13 |
|
| 14 |
+
This model is a fine-tuned version of [Qwen/Qwen2.5-0.5B-Instruct](https://huggingface.co/Qwen/Qwen2.5-0.5B-Instruct).
|
| 15 |
It has been trained using [TRL](https://github.com/huggingface/trl).
|
| 16 |
|
| 17 |
## Quick start
|
|
|
|
| 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="YangZhoumill/Qwen2.5-1.5B-Open-R1-Distill", 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/stevenzhou0816100/huggingface/runs/edxcy052)
|
| 31 |
|
| 32 |
|
| 33 |
This model was trained with SFT.
|
all_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"total_flos":
|
| 3 |
-
"train_loss": 0.
|
| 4 |
-
"train_runtime":
|
| 5 |
-
"train_samples":
|
| 6 |
-
"train_samples_per_second": 0.
|
| 7 |
-
"train_steps_per_second": 0.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"total_flos": 4.404961752383488e+16,
|
| 3 |
+
"train_loss": 0.28405343046555154,
|
| 4 |
+
"train_runtime": 1315.0218,
|
| 5 |
+
"train_samples": 10016,
|
| 6 |
+
"train_samples_per_second": 0.989,
|
| 7 |
+
"train_steps_per_second": 0.989
|
| 8 |
}
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"total_flos":
|
| 3 |
-
"train_loss": 0.
|
| 4 |
-
"train_runtime":
|
| 5 |
-
"train_samples":
|
| 6 |
-
"train_samples_per_second": 0.
|
| 7 |
-
"train_steps_per_second": 0.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"total_flos": 4.404961752383488e+16,
|
| 3 |
+
"train_loss": 0.28405343046555154,
|
| 4 |
+
"train_runtime": 1315.0218,
|
| 5 |
+
"train_samples": 10016,
|
| 6 |
+
"train_samples_per_second": 0.989,
|
| 7 |
+
"train_steps_per_second": 0.989
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -3,86 +3,1843 @@
|
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
"epoch": 1.0,
|
| 5 |
"eval_steps": 500,
|
| 6 |
-
"global_step":
|
| 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.
|
| 13 |
-
"grad_norm":
|
| 14 |
-
"learning_rate":
|
| 15 |
-
"loss":
|
| 16 |
"step": 5
|
| 17 |
},
|
| 18 |
{
|
| 19 |
-
"epoch": 0.
|
| 20 |
-
"grad_norm":
|
| 21 |
-
"learning_rate":
|
| 22 |
-
"loss":
|
| 23 |
"step": 10
|
| 24 |
},
|
| 25 |
{
|
| 26 |
-
"epoch": 0.
|
| 27 |
-
"grad_norm":
|
| 28 |
-
"learning_rate":
|
| 29 |
-
"loss": 0.
|
| 30 |
"step": 15
|
| 31 |
},
|
| 32 |
{
|
| 33 |
-
"epoch": 0.
|
| 34 |
-
"grad_norm":
|
| 35 |
-
"learning_rate":
|
| 36 |
-
"loss": 0.
|
| 37 |
"step": 20
|
| 38 |
},
|
| 39 |
{
|
| 40 |
-
"epoch": 0.
|
| 41 |
-
"grad_norm":
|
| 42 |
-
"learning_rate":
|
| 43 |
-
"loss": 0.
|
| 44 |
"step": 25
|
| 45 |
},
|
| 46 |
{
|
| 47 |
-
"epoch": 0.
|
| 48 |
-
"grad_norm":
|
| 49 |
-
"learning_rate": 2.
|
| 50 |
-
"loss": 0.
|
| 51 |
"step": 30
|
| 52 |
},
|
| 53 |
{
|
| 54 |
-
"epoch": 0.
|
| 55 |
-
"grad_norm":
|
| 56 |
-
"learning_rate":
|
| 57 |
-
"loss": 0.
|
| 58 |
"step": 35
|
| 59 |
},
|
| 60 |
{
|
| 61 |
-
"epoch": 0.
|
| 62 |
-
"grad_norm":
|
| 63 |
-
"learning_rate":
|
| 64 |
-
"loss": 0.
|
| 65 |
"step": 40
|
| 66 |
},
|
| 67 |
{
|
| 68 |
-
"epoch": 0.
|
| 69 |
-
"grad_norm":
|
| 70 |
-
"learning_rate":
|
| 71 |
-
"loss": 0.
|
| 72 |
"step": 45
|
| 73 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 74 |
{
|
| 75 |
"epoch": 1.0,
|
| 76 |
-
"step":
|
| 77 |
-
"total_flos":
|
| 78 |
-
"train_loss": 0.
|
| 79 |
-
"train_runtime":
|
| 80 |
-
"train_samples_per_second": 0.
|
| 81 |
-
"train_steps_per_second": 0.
|
| 82 |
}
|
| 83 |
],
|
| 84 |
"logging_steps": 5,
|
| 85 |
-
"max_steps":
|
| 86 |
"num_input_tokens_seen": 0,
|
| 87 |
"num_train_epochs": 1,
|
| 88 |
"save_steps": 100,
|
|
@@ -98,7 +1855,7 @@
|
|
| 98 |
"attributes": {}
|
| 99 |
}
|
| 100 |
},
|
| 101 |
-
"total_flos":
|
| 102 |
"train_batch_size": 1,
|
| 103 |
"trial_name": null,
|
| 104 |
"trial_params": null
|
|
|
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
"epoch": 1.0,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 1300,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
+
"epoch": 0.0038461538461538464,
|
| 13 |
+
"grad_norm": 68.5,
|
| 14 |
+
"learning_rate": 3.846153846153847e-06,
|
| 15 |
+
"loss": 1.442,
|
| 16 |
"step": 5
|
| 17 |
},
|
| 18 |
{
|
| 19 |
+
"epoch": 0.007692307692307693,
|
| 20 |
+
"grad_norm": 28.875,
|
| 21 |
+
"learning_rate": 7.692307692307694e-06,
|
| 22 |
+
"loss": 1.2205,
|
| 23 |
"step": 10
|
| 24 |
},
|
| 25 |
{
|
| 26 |
+
"epoch": 0.011538461538461539,
|
| 27 |
+
"grad_norm": 14.6875,
|
| 28 |
+
"learning_rate": 1.153846153846154e-05,
|
| 29 |
+
"loss": 0.9428,
|
| 30 |
"step": 15
|
| 31 |
},
|
| 32 |
{
|
| 33 |
+
"epoch": 0.015384615384615385,
|
| 34 |
+
"grad_norm": 9.375,
|
| 35 |
+
"learning_rate": 1.5384615384615387e-05,
|
| 36 |
+
"loss": 0.7695,
|
| 37 |
"step": 20
|
| 38 |
},
|
| 39 |
{
|
| 40 |
+
"epoch": 0.019230769230769232,
|
| 41 |
+
"grad_norm": 14528.0,
|
| 42 |
+
"learning_rate": 1.923076923076923e-05,
|
| 43 |
+
"loss": 0.7043,
|
| 44 |
"step": 25
|
| 45 |
},
|
| 46 |
{
|
| 47 |
+
"epoch": 0.023076923076923078,
|
| 48 |
+
"grad_norm": 4.21875,
|
| 49 |
+
"learning_rate": 2.307692307692308e-05,
|
| 50 |
+
"loss": 0.6085,
|
| 51 |
"step": 30
|
| 52 |
},
|
| 53 |
{
|
| 54 |
+
"epoch": 0.026923076923076925,
|
| 55 |
+
"grad_norm": 7.40625,
|
| 56 |
+
"learning_rate": 2.6923076923076923e-05,
|
| 57 |
+
"loss": 0.575,
|
| 58 |
"step": 35
|
| 59 |
},
|
| 60 |
{
|
| 61 |
+
"epoch": 0.03076923076923077,
|
| 62 |
+
"grad_norm": 5.40625,
|
| 63 |
+
"learning_rate": 3.0769230769230774e-05,
|
| 64 |
+
"loss": 0.5192,
|
| 65 |
"step": 40
|
| 66 |
},
|
| 67 |
{
|
| 68 |
+
"epoch": 0.03461538461538462,
|
| 69 |
+
"grad_norm": 5.53125,
|
| 70 |
+
"learning_rate": 3.461538461538462e-05,
|
| 71 |
+
"loss": 0.5032,
|
| 72 |
"step": 45
|
| 73 |
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.038461538461538464,
|
| 76 |
+
"grad_norm": 6.09375,
|
| 77 |
+
"learning_rate": 3.846153846153846e-05,
|
| 78 |
+
"loss": 0.4601,
|
| 79 |
+
"step": 50
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.04230769230769231,
|
| 83 |
+
"grad_norm": 3.9375,
|
| 84 |
+
"learning_rate": 4.230769230769231e-05,
|
| 85 |
+
"loss": 0.4175,
|
| 86 |
+
"step": 55
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.046153846153846156,
|
| 90 |
+
"grad_norm": 4.8125,
|
| 91 |
+
"learning_rate": 4.615384615384616e-05,
|
| 92 |
+
"loss": 0.4093,
|
| 93 |
+
"step": 60
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.05,
|
| 97 |
+
"grad_norm": 4.9375,
|
| 98 |
+
"learning_rate": 5e-05,
|
| 99 |
+
"loss": 0.387,
|
| 100 |
+
"step": 65
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.05384615384615385,
|
| 104 |
+
"grad_norm": 4.78125,
|
| 105 |
+
"learning_rate": 4.99981800791248e-05,
|
| 106 |
+
"loss": 0.4415,
|
| 107 |
+
"step": 70
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.057692307692307696,
|
| 111 |
+
"grad_norm": 5.0,
|
| 112 |
+
"learning_rate": 4.9992720610909145e-05,
|
| 113 |
+
"loss": 0.4325,
|
| 114 |
+
"step": 75
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.06153846153846154,
|
| 118 |
+
"grad_norm": 4.15625,
|
| 119 |
+
"learning_rate": 4.998362247853528e-05,
|
| 120 |
+
"loss": 0.3797,
|
| 121 |
+
"step": 80
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.06538461538461539,
|
| 125 |
+
"grad_norm": 5.28125,
|
| 126 |
+
"learning_rate": 4.997088715381485e-05,
|
| 127 |
+
"loss": 0.3574,
|
| 128 |
+
"step": 85
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.06923076923076923,
|
| 132 |
+
"grad_norm": 3.78125,
|
| 133 |
+
"learning_rate": 4.995451669695081e-05,
|
| 134 |
+
"loss": 0.3803,
|
| 135 |
+
"step": 90
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.07307692307692308,
|
| 139 |
+
"grad_norm": 3.15625,
|
| 140 |
+
"learning_rate": 4.993451375620418e-05,
|
| 141 |
+
"loss": 0.3751,
|
| 142 |
+
"step": 95
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.07692307692307693,
|
| 146 |
+
"grad_norm": 2.640625,
|
| 147 |
+
"learning_rate": 4.991088156746554e-05,
|
| 148 |
+
"loss": 0.3477,
|
| 149 |
+
"step": 100
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.08076923076923077,
|
| 153 |
+
"grad_norm": 2.4375,
|
| 154 |
+
"learning_rate": 4.988362395373167e-05,
|
| 155 |
+
"loss": 0.2947,
|
| 156 |
+
"step": 105
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.08461538461538462,
|
| 160 |
+
"grad_norm": 4.21875,
|
| 161 |
+
"learning_rate": 4.985274532448704e-05,
|
| 162 |
+
"loss": 0.3417,
|
| 163 |
+
"step": 110
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.08846153846153847,
|
| 167 |
+
"grad_norm": 6.28125,
|
| 168 |
+
"learning_rate": 4.981825067499049e-05,
|
| 169 |
+
"loss": 0.3585,
|
| 170 |
+
"step": 115
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.09230769230769231,
|
| 174 |
+
"grad_norm": 2.921875,
|
| 175 |
+
"learning_rate": 4.978014558546715e-05,
|
| 176 |
+
"loss": 0.3465,
|
| 177 |
+
"step": 120
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.09615384615384616,
|
| 181 |
+
"grad_norm": 3.25,
|
| 182 |
+
"learning_rate": 4.973843622020571e-05,
|
| 183 |
+
"loss": 0.2972,
|
| 184 |
+
"step": 125
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.1,
|
| 188 |
+
"grad_norm": 4.65625,
|
| 189 |
+
"learning_rate": 4.9693129326561254e-05,
|
| 190 |
+
"loss": 0.3853,
|
| 191 |
+
"step": 130
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.10384615384615385,
|
| 195 |
+
"grad_norm": 4.21875,
|
| 196 |
+
"learning_rate": 4.9644232233863685e-05,
|
| 197 |
+
"loss": 0.3619,
|
| 198 |
+
"step": 135
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.1076923076923077,
|
| 202 |
+
"grad_norm": 3.921875,
|
| 203 |
+
"learning_rate": 4.95917528522321e-05,
|
| 204 |
+
"loss": 0.3283,
|
| 205 |
+
"step": 140
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.11153846153846154,
|
| 209 |
+
"grad_norm": 2.703125,
|
| 210 |
+
"learning_rate": 4.953569967129513e-05,
|
| 211 |
+
"loss": 0.368,
|
| 212 |
+
"step": 145
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.11538461538461539,
|
| 216 |
+
"grad_norm": 3.375,
|
| 217 |
+
"learning_rate": 4.947608175881757e-05,
|
| 218 |
+
"loss": 0.3791,
|
| 219 |
+
"step": 150
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.11923076923076924,
|
| 223 |
+
"grad_norm": 3.65625,
|
| 224 |
+
"learning_rate": 4.9412908759233535e-05,
|
| 225 |
+
"loss": 0.3163,
|
| 226 |
+
"step": 155
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.12307692307692308,
|
| 230 |
+
"grad_norm": 5.625,
|
| 231 |
+
"learning_rate": 4.9346190892086174e-05,
|
| 232 |
+
"loss": 0.3239,
|
| 233 |
+
"step": 160
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.12692307692307692,
|
| 237 |
+
"grad_norm": 3.484375,
|
| 238 |
+
"learning_rate": 4.927593895037453e-05,
|
| 239 |
+
"loss": 0.2991,
|
| 240 |
+
"step": 165
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.13076923076923078,
|
| 244 |
+
"grad_norm": 3.421875,
|
| 245 |
+
"learning_rate": 4.920216429880752e-05,
|
| 246 |
+
"loss": 0.3416,
|
| 247 |
+
"step": 170
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.1346153846153846,
|
| 251 |
+
"grad_norm": 2.546875,
|
| 252 |
+
"learning_rate": 4.912487887196544e-05,
|
| 253 |
+
"loss": 0.3039,
|
| 254 |
+
"step": 175
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.13846153846153847,
|
| 258 |
+
"grad_norm": 3.25,
|
| 259 |
+
"learning_rate": 4.904409517236934e-05,
|
| 260 |
+
"loss": 0.2838,
|
| 261 |
+
"step": 180
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.1423076923076923,
|
| 265 |
+
"grad_norm": 3.0625,
|
| 266 |
+
"learning_rate": 4.895982626845843e-05,
|
| 267 |
+
"loss": 0.3172,
|
| 268 |
+
"step": 185
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.14615384615384616,
|
| 272 |
+
"grad_norm": 2.546875,
|
| 273 |
+
"learning_rate": 4.887208579247604e-05,
|
| 274 |
+
"loss": 0.2793,
|
| 275 |
+
"step": 190
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.15,
|
| 279 |
+
"grad_norm": 3.8125,
|
| 280 |
+
"learning_rate": 4.878088793826428e-05,
|
| 281 |
+
"loss": 0.3016,
|
| 282 |
+
"step": 195
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.15384615384615385,
|
| 286 |
+
"grad_norm": 2.984375,
|
| 287 |
+
"learning_rate": 4.868624745896793e-05,
|
| 288 |
+
"loss": 0.2968,
|
| 289 |
+
"step": 200
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.1576923076923077,
|
| 293 |
+
"grad_norm": 3.75,
|
| 294 |
+
"learning_rate": 4.858817966464777e-05,
|
| 295 |
+
"loss": 0.295,
|
| 296 |
+
"step": 205
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.16153846153846155,
|
| 300 |
+
"grad_norm": 2.0625,
|
| 301 |
+
"learning_rate": 4.848670041980392e-05,
|
| 302 |
+
"loss": 0.3002,
|
| 303 |
+
"step": 210
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.16538461538461538,
|
| 307 |
+
"grad_norm": 2.671875,
|
| 308 |
+
"learning_rate": 4.838182614080934e-05,
|
| 309 |
+
"loss": 0.3005,
|
| 310 |
+
"step": 215
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.16923076923076924,
|
| 314 |
+
"grad_norm": 2.21875,
|
| 315 |
+
"learning_rate": 4.827357379325422e-05,
|
| 316 |
+
"loss": 0.2546,
|
| 317 |
+
"step": 220
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.17307692307692307,
|
| 321 |
+
"grad_norm": 2.140625,
|
| 322 |
+
"learning_rate": 4.816196088920143e-05,
|
| 323 |
+
"loss": 0.3147,
|
| 324 |
+
"step": 225
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.17692307692307693,
|
| 328 |
+
"grad_norm": 2.34375,
|
| 329 |
+
"learning_rate": 4.804700548435353e-05,
|
| 330 |
+
"loss": 0.276,
|
| 331 |
+
"step": 230
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.18076923076923077,
|
| 335 |
+
"grad_norm": 2.828125,
|
| 336 |
+
"learning_rate": 4.7928726175131955e-05,
|
| 337 |
+
"loss": 0.3404,
|
| 338 |
+
"step": 235
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.18461538461538463,
|
| 342 |
+
"grad_norm": 3.65625,
|
| 343 |
+
"learning_rate": 4.780714209566861e-05,
|
| 344 |
+
"loss": 0.3394,
|
| 345 |
+
"step": 240
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.18846153846153846,
|
| 349 |
+
"grad_norm": 2.3125,
|
| 350 |
+
"learning_rate": 4.768227291471055e-05,
|
| 351 |
+
"loss": 0.2997,
|
| 352 |
+
"step": 245
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.19230769230769232,
|
| 356 |
+
"grad_norm": 2.765625,
|
| 357 |
+
"learning_rate": 4.755413883243812e-05,
|
| 358 |
+
"loss": 0.2843,
|
| 359 |
+
"step": 250
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.19615384615384615,
|
| 363 |
+
"grad_norm": 2.828125,
|
| 364 |
+
"learning_rate": 4.7422760577197226e-05,
|
| 365 |
+
"loss": 0.2963,
|
| 366 |
+
"step": 255
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.2,
|
| 370 |
+
"grad_norm": 2.765625,
|
| 371 |
+
"learning_rate": 4.7288159402146e-05,
|
| 372 |
+
"loss": 0.2854,
|
| 373 |
+
"step": 260
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.20384615384615384,
|
| 377 |
+
"grad_norm": 2.859375,
|
| 378 |
+
"learning_rate": 4.715035708181676e-05,
|
| 379 |
+
"loss": 0.3067,
|
| 380 |
+
"step": 265
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.2076923076923077,
|
| 384 |
+
"grad_norm": 2.6875,
|
| 385 |
+
"learning_rate": 4.7009375908593434e-05,
|
| 386 |
+
"loss": 0.2886,
|
| 387 |
+
"step": 270
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.21153846153846154,
|
| 391 |
+
"grad_norm": 2.765625,
|
| 392 |
+
"learning_rate": 4.6865238689105386e-05,
|
| 393 |
+
"loss": 0.3067,
|
| 394 |
+
"step": 275
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.2153846153846154,
|
| 398 |
+
"grad_norm": 3.171875,
|
| 399 |
+
"learning_rate": 4.67179687405379e-05,
|
| 400 |
+
"loss": 0.3321,
|
| 401 |
+
"step": 280
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.21923076923076923,
|
| 405 |
+
"grad_norm": 2.515625,
|
| 406 |
+
"learning_rate": 4.65675898868602e-05,
|
| 407 |
+
"loss": 0.283,
|
| 408 |
+
"step": 285
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.2230769230769231,
|
| 412 |
+
"grad_norm": 1.8203125,
|
| 413 |
+
"learning_rate": 4.6414126454971406e-05,
|
| 414 |
+
"loss": 0.2626,
|
| 415 |
+
"step": 290
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.22692307692307692,
|
| 419 |
+
"grad_norm": 2.171875,
|
| 420 |
+
"learning_rate": 4.625760327076513e-05,
|
| 421 |
+
"loss": 0.2789,
|
| 422 |
+
"step": 295
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.23076923076923078,
|
| 426 |
+
"grad_norm": 2.65625,
|
| 427 |
+
"learning_rate": 4.6098045655113416e-05,
|
| 428 |
+
"loss": 0.3019,
|
| 429 |
+
"step": 300
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.23461538461538461,
|
| 433 |
+
"grad_norm": 1.765625,
|
| 434 |
+
"learning_rate": 4.5935479419770525e-05,
|
| 435 |
+
"loss": 0.2696,
|
| 436 |
+
"step": 305
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.23846153846153847,
|
| 440 |
+
"grad_norm": 2.8125,
|
| 441 |
+
"learning_rate": 4.576993086319737e-05,
|
| 442 |
+
"loss": 0.2759,
|
| 443 |
+
"step": 310
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.2423076923076923,
|
| 447 |
+
"grad_norm": 2.671875,
|
| 448 |
+
"learning_rate": 4.56014267663072e-05,
|
| 449 |
+
"loss": 0.2688,
|
| 450 |
+
"step": 315
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.24615384615384617,
|
| 454 |
+
"grad_norm": 1.953125,
|
| 455 |
+
"learning_rate": 4.542999438813322e-05,
|
| 456 |
+
"loss": 0.2696,
|
| 457 |
+
"step": 320
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.25,
|
| 461 |
+
"grad_norm": 2.40625,
|
| 462 |
+
"learning_rate": 4.5255661461418854e-05,
|
| 463 |
+
"loss": 0.2571,
|
| 464 |
+
"step": 325
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.25384615384615383,
|
| 468 |
+
"grad_norm": 1.9296875,
|
| 469 |
+
"learning_rate": 4.5078456188131464e-05,
|
| 470 |
+
"loss": 0.2871,
|
| 471 |
+
"step": 330
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.25769230769230766,
|
| 475 |
+
"grad_norm": 1.9453125,
|
| 476 |
+
"learning_rate": 4.4898407234900006e-05,
|
| 477 |
+
"loss": 0.2977,
|
| 478 |
+
"step": 335
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.26153846153846155,
|
| 482 |
+
"grad_norm": 1.8203125,
|
| 483 |
+
"learning_rate": 4.4715543728377706e-05,
|
| 484 |
+
"loss": 0.2467,
|
| 485 |
+
"step": 340
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.2653846153846154,
|
| 489 |
+
"grad_norm": 2.703125,
|
| 490 |
+
"learning_rate": 4.452989525053014e-05,
|
| 491 |
+
"loss": 0.2563,
|
| 492 |
+
"step": 345
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.2692307692307692,
|
| 496 |
+
"grad_norm": 2.46875,
|
| 497 |
+
"learning_rate": 4.434149183384977e-05,
|
| 498 |
+
"loss": 0.31,
|
| 499 |
+
"step": 350
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.27307692307692305,
|
| 503 |
+
"grad_norm": 2.078125,
|
| 504 |
+
"learning_rate": 4.41503639564976e-05,
|
| 505 |
+
"loss": 0.25,
|
| 506 |
+
"step": 355
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.27692307692307694,
|
| 510 |
+
"grad_norm": 2.453125,
|
| 511 |
+
"learning_rate": 4.3956542537372596e-05,
|
| 512 |
+
"loss": 0.2497,
|
| 513 |
+
"step": 360
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.28076923076923077,
|
| 517 |
+
"grad_norm": 2.53125,
|
| 518 |
+
"learning_rate": 4.3760058931110054e-05,
|
| 519 |
+
"loss": 0.2672,
|
| 520 |
+
"step": 365
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.2846153846153846,
|
| 524 |
+
"grad_norm": 1.9921875,
|
| 525 |
+
"learning_rate": 4.356094492300922e-05,
|
| 526 |
+
"loss": 0.2678,
|
| 527 |
+
"step": 370
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.28846153846153844,
|
| 531 |
+
"grad_norm": 3.421875,
|
| 532 |
+
"learning_rate": 4.3359232723891423e-05,
|
| 533 |
+
"loss": 0.2926,
|
| 534 |
+
"step": 375
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.2923076923076923,
|
| 538 |
+
"grad_norm": 2.40625,
|
| 539 |
+
"learning_rate": 4.315495496488927e-05,
|
| 540 |
+
"loss": 0.3105,
|
| 541 |
+
"step": 380
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.29615384615384616,
|
| 545 |
+
"grad_norm": 2.171875,
|
| 546 |
+
"learning_rate": 4.294814469216794e-05,
|
| 547 |
+
"loss": 0.2627,
|
| 548 |
+
"step": 385
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.3,
|
| 552 |
+
"grad_norm": 1.96875,
|
| 553 |
+
"learning_rate": 4.2738835361579175e-05,
|
| 554 |
+
"loss": 0.257,
|
| 555 |
+
"step": 390
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.3038461538461538,
|
| 559 |
+
"grad_norm": 3.78125,
|
| 560 |
+
"learning_rate": 4.252706083324923e-05,
|
| 561 |
+
"loss": 0.3106,
|
| 562 |
+
"step": 395
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.3076923076923077,
|
| 566 |
+
"grad_norm": 2.859375,
|
| 567 |
+
"learning_rate": 4.2312855366101216e-05,
|
| 568 |
+
"loss": 0.2824,
|
| 569 |
+
"step": 400
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.31153846153846154,
|
| 573 |
+
"grad_norm": 3.359375,
|
| 574 |
+
"learning_rate": 4.2096253612313006e-05,
|
| 575 |
+
"loss": 0.2902,
|
| 576 |
+
"step": 405
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.3153846153846154,
|
| 580 |
+
"grad_norm": 2.328125,
|
| 581 |
+
"learning_rate": 4.187729061171158e-05,
|
| 582 |
+
"loss": 0.263,
|
| 583 |
+
"step": 410
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.3192307692307692,
|
| 587 |
+
"grad_norm": 1.59375,
|
| 588 |
+
"learning_rate": 4.165600178610454e-05,
|
| 589 |
+
"loss": 0.2842,
|
| 590 |
+
"step": 415
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.3230769230769231,
|
| 594 |
+
"grad_norm": 2.5625,
|
| 595 |
+
"learning_rate": 4.143242293354995e-05,
|
| 596 |
+
"loss": 0.2506,
|
| 597 |
+
"step": 420
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.3269230769230769,
|
| 601 |
+
"grad_norm": 2.15625,
|
| 602 |
+
"learning_rate": 4.120659022256523e-05,
|
| 603 |
+
"loss": 0.3082,
|
| 604 |
+
"step": 425
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.33076923076923076,
|
| 608 |
+
"grad_norm": 2.09375,
|
| 609 |
+
"learning_rate": 4.097854018627616e-05,
|
| 610 |
+
"loss": 0.3015,
|
| 611 |
+
"step": 430
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.3346153846153846,
|
| 615 |
+
"grad_norm": 2.765625,
|
| 616 |
+
"learning_rate": 4.074830971650691e-05,
|
| 617 |
+
"loss": 0.2668,
|
| 618 |
+
"step": 435
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.3384615384615385,
|
| 622 |
+
"grad_norm": 3.0,
|
| 623 |
+
"learning_rate": 4.0515936057811924e-05,
|
| 624 |
+
"loss": 0.2663,
|
| 625 |
+
"step": 440
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.3423076923076923,
|
| 629 |
+
"grad_norm": 1.40625,
|
| 630 |
+
"learning_rate": 4.0281456801451004e-05,
|
| 631 |
+
"loss": 0.2735,
|
| 632 |
+
"step": 445
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.34615384615384615,
|
| 636 |
+
"grad_norm": 2.25,
|
| 637 |
+
"learning_rate": 4.0044909879308004e-05,
|
| 638 |
+
"loss": 0.2879,
|
| 639 |
+
"step": 450
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.35,
|
| 643 |
+
"grad_norm": 2.265625,
|
| 644 |
+
"learning_rate": 3.980633355775461e-05,
|
| 645 |
+
"loss": 0.2408,
|
| 646 |
+
"step": 455
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.35384615384615387,
|
| 650 |
+
"grad_norm": 2.375,
|
| 651 |
+
"learning_rate": 3.956576643145997e-05,
|
| 652 |
+
"loss": 0.2576,
|
| 653 |
+
"step": 460
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.3576923076923077,
|
| 657 |
+
"grad_norm": 2.3125,
|
| 658 |
+
"learning_rate": 3.9323247417147204e-05,
|
| 659 |
+
"loss": 0.263,
|
| 660 |
+
"step": 465
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.36153846153846153,
|
| 664 |
+
"grad_norm": 1.796875,
|
| 665 |
+
"learning_rate": 3.907881574729781e-05,
|
| 666 |
+
"loss": 0.2473,
|
| 667 |
+
"step": 470
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.36538461538461536,
|
| 671 |
+
"grad_norm": 1.90625,
|
| 672 |
+
"learning_rate": 3.8832510963804976e-05,
|
| 673 |
+
"loss": 0.2811,
|
| 674 |
+
"step": 475
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.36923076923076925,
|
| 678 |
+
"grad_norm": 2.671875,
|
| 679 |
+
"learning_rate": 3.858437291157689e-05,
|
| 680 |
+
"loss": 0.3075,
|
| 681 |
+
"step": 480
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.3730769230769231,
|
| 685 |
+
"grad_norm": 1.9921875,
|
| 686 |
+
"learning_rate": 3.833444173209102e-05,
|
| 687 |
+
"loss": 0.2493,
|
| 688 |
+
"step": 485
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.3769230769230769,
|
| 692 |
+
"grad_norm": 1.9453125,
|
| 693 |
+
"learning_rate": 3.80827578569003e-05,
|
| 694 |
+
"loss": 0.2414,
|
| 695 |
+
"step": 490
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.38076923076923075,
|
| 699 |
+
"grad_norm": 1.65625,
|
| 700 |
+
"learning_rate": 3.782936200109261e-05,
|
| 701 |
+
"loss": 0.2632,
|
| 702 |
+
"step": 495
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.38461538461538464,
|
| 706 |
+
"grad_norm": 1.8671875,
|
| 707 |
+
"learning_rate": 3.7574295156704175e-05,
|
| 708 |
+
"loss": 0.2756,
|
| 709 |
+
"step": 500
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.38846153846153847,
|
| 713 |
+
"grad_norm": 2.015625,
|
| 714 |
+
"learning_rate": 3.731759858608832e-05,
|
| 715 |
+
"loss": 0.2684,
|
| 716 |
+
"step": 505
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.3923076923076923,
|
| 720 |
+
"grad_norm": 2.375,
|
| 721 |
+
"learning_rate": 3.705931381524034e-05,
|
| 722 |
+
"loss": 0.2528,
|
| 723 |
+
"step": 510
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.39615384615384613,
|
| 727 |
+
"grad_norm": 1.6640625,
|
| 728 |
+
"learning_rate": 3.679948262707993e-05,
|
| 729 |
+
"loss": 0.2669,
|
| 730 |
+
"step": 515
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.4,
|
| 734 |
+
"grad_norm": 2.078125,
|
| 735 |
+
"learning_rate": 3.6538147054691817e-05,
|
| 736 |
+
"loss": 0.2469,
|
| 737 |
+
"step": 520
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.40384615384615385,
|
| 741 |
+
"grad_norm": 2.28125,
|
| 742 |
+
"learning_rate": 3.62753493745261e-05,
|
| 743 |
+
"loss": 0.251,
|
| 744 |
+
"step": 525
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.4076923076923077,
|
| 748 |
+
"grad_norm": 1.9375,
|
| 749 |
+
"learning_rate": 3.601113209955916e-05,
|
| 750 |
+
"loss": 0.2568,
|
| 751 |
+
"step": 530
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.4115384615384615,
|
| 755 |
+
"grad_norm": 2.015625,
|
| 756 |
+
"learning_rate": 3.574553797241625e-05,
|
| 757 |
+
"loss": 0.2698,
|
| 758 |
+
"step": 535
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.4153846153846154,
|
| 762 |
+
"grad_norm": 2.109375,
|
| 763 |
+
"learning_rate": 3.5478609958457056e-05,
|
| 764 |
+
"loss": 0.2517,
|
| 765 |
+
"step": 540
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.41923076923076924,
|
| 769 |
+
"grad_norm": 2.890625,
|
| 770 |
+
"learning_rate": 3.52103912388251e-05,
|
| 771 |
+
"loss": 0.2793,
|
| 772 |
+
"step": 545
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.4230769230769231,
|
| 776 |
+
"grad_norm": 2.5,
|
| 777 |
+
"learning_rate": 3.4940925203462347e-05,
|
| 778 |
+
"loss": 0.244,
|
| 779 |
+
"step": 550
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.4269230769230769,
|
| 783 |
+
"grad_norm": 1.953125,
|
| 784 |
+
"learning_rate": 3.467025544408994e-05,
|
| 785 |
+
"loss": 0.2511,
|
| 786 |
+
"step": 555
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.4307692307692308,
|
| 790 |
+
"grad_norm": 3.15625,
|
| 791 |
+
"learning_rate": 3.4398425747156346e-05,
|
| 792 |
+
"loss": 0.2579,
|
| 793 |
+
"step": 560
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.4346153846153846,
|
| 797 |
+
"grad_norm": 1.921875,
|
| 798 |
+
"learning_rate": 3.412548008675403e-05,
|
| 799 |
+
"loss": 0.2595,
|
| 800 |
+
"step": 565
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.43846153846153846,
|
| 804 |
+
"grad_norm": 1.9296875,
|
| 805 |
+
"learning_rate": 3.3851462617505655e-05,
|
| 806 |
+
"loss": 0.2263,
|
| 807 |
+
"step": 570
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.4423076923076923,
|
| 811 |
+
"grad_norm": 1.671875,
|
| 812 |
+
"learning_rate": 3.357641766742121e-05,
|
| 813 |
+
"loss": 0.2641,
|
| 814 |
+
"step": 575
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.4461538461538462,
|
| 818 |
+
"grad_norm": 1.5390625,
|
| 819 |
+
"learning_rate": 3.3300389730727064e-05,
|
| 820 |
+
"loss": 0.2693,
|
| 821 |
+
"step": 580
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.45,
|
| 825 |
+
"grad_norm": 1.4765625,
|
| 826 |
+
"learning_rate": 3.3023423460667985e-05,
|
| 827 |
+
"loss": 0.2478,
|
| 828 |
+
"step": 585
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.45384615384615384,
|
| 832 |
+
"grad_norm": 1.890625,
|
| 833 |
+
"learning_rate": 3.27455636622837e-05,
|
| 834 |
+
"loss": 0.24,
|
| 835 |
+
"step": 590
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.4576923076923077,
|
| 839 |
+
"grad_norm": 2.078125,
|
| 840 |
+
"learning_rate": 3.2466855285160644e-05,
|
| 841 |
+
"loss": 0.282,
|
| 842 |
+
"step": 595
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.46153846153846156,
|
| 846 |
+
"grad_norm": 2.4375,
|
| 847 |
+
"learning_rate": 3.218734341616047e-05,
|
| 848 |
+
"loss": 0.2595,
|
| 849 |
+
"step": 600
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.4653846153846154,
|
| 853 |
+
"grad_norm": 1.859375,
|
| 854 |
+
"learning_rate": 3.1907073272126306e-05,
|
| 855 |
+
"loss": 0.2573,
|
| 856 |
+
"step": 605
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.46923076923076923,
|
| 860 |
+
"grad_norm": 2.296875,
|
| 861 |
+
"learning_rate": 3.162609019256802e-05,
|
| 862 |
+
"loss": 0.2668,
|
| 863 |
+
"step": 610
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.47307692307692306,
|
| 867 |
+
"grad_norm": 1.6640625,
|
| 868 |
+
"learning_rate": 3.134443963232755e-05,
|
| 869 |
+
"loss": 0.2264,
|
| 870 |
+
"step": 615
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.47692307692307695,
|
| 874 |
+
"grad_norm": 2.125,
|
| 875 |
+
"learning_rate": 3.1062167154225725e-05,
|
| 876 |
+
"loss": 0.259,
|
| 877 |
+
"step": 620
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.4807692307692308,
|
| 881 |
+
"grad_norm": 2.625,
|
| 882 |
+
"learning_rate": 3.0779318421691464e-05,
|
| 883 |
+
"loss": 0.2848,
|
| 884 |
+
"step": 625
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.4846153846153846,
|
| 888 |
+
"grad_norm": 1.9921875,
|
| 889 |
+
"learning_rate": 3.0495939191374796e-05,
|
| 890 |
+
"loss": 0.2425,
|
| 891 |
+
"step": 630
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.48846153846153845,
|
| 895 |
+
"grad_norm": 1.6171875,
|
| 896 |
+
"learning_rate": 3.021207530574477e-05,
|
| 897 |
+
"loss": 0.2488,
|
| 898 |
+
"step": 635
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.49230769230769234,
|
| 902 |
+
"grad_norm": 2.015625,
|
| 903 |
+
"learning_rate": 2.9927772685673505e-05,
|
| 904 |
+
"loss": 0.2589,
|
| 905 |
+
"step": 640
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.49615384615384617,
|
| 909 |
+
"grad_norm": 1.9921875,
|
| 910 |
+
"learning_rate": 2.9643077323007484e-05,
|
| 911 |
+
"loss": 0.249,
|
| 912 |
+
"step": 645
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.5,
|
| 916 |
+
"grad_norm": 1.75,
|
| 917 |
+
"learning_rate": 2.9358035273127483e-05,
|
| 918 |
+
"loss": 0.209,
|
| 919 |
+
"step": 650
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.5038461538461538,
|
| 923 |
+
"grad_norm": 1.828125,
|
| 924 |
+
"learning_rate": 2.9072692647498106e-05,
|
| 925 |
+
"loss": 0.2399,
|
| 926 |
+
"step": 655
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.5076923076923077,
|
| 930 |
+
"grad_norm": 1.4375,
|
| 931 |
+
"learning_rate": 2.8787095606208335e-05,
|
| 932 |
+
"loss": 0.1993,
|
| 933 |
+
"step": 660
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.5115384615384615,
|
| 937 |
+
"grad_norm": 1.9453125,
|
| 938 |
+
"learning_rate": 2.8501290350504162e-05,
|
| 939 |
+
"loss": 0.2196,
|
| 940 |
+
"step": 665
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.5153846153846153,
|
| 944 |
+
"grad_norm": 1.921875,
|
| 945 |
+
"learning_rate": 2.821532311531455e-05,
|
| 946 |
+
"loss": 0.2367,
|
| 947 |
+
"step": 670
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.5192307692307693,
|
| 951 |
+
"grad_norm": 1.8046875,
|
| 952 |
+
"learning_rate": 2.792924016177205e-05,
|
| 953 |
+
"loss": 0.2782,
|
| 954 |
+
"step": 675
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.5230769230769231,
|
| 958 |
+
"grad_norm": 1.296875,
|
| 959 |
+
"learning_rate": 2.764308776972903e-05,
|
| 960 |
+
"loss": 0.2169,
|
| 961 |
+
"step": 680
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.5269230769230769,
|
| 965 |
+
"grad_norm": 1.921875,
|
| 966 |
+
"learning_rate": 2.735691223027098e-05,
|
| 967 |
+
"loss": 0.222,
|
| 968 |
+
"step": 685
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.5307692307692308,
|
| 972 |
+
"grad_norm": 1.984375,
|
| 973 |
+
"learning_rate": 2.7070759838227955e-05,
|
| 974 |
+
"loss": 0.2558,
|
| 975 |
+
"step": 690
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.5346153846153846,
|
| 979 |
+
"grad_norm": 2.03125,
|
| 980 |
+
"learning_rate": 2.678467688468545e-05,
|
| 981 |
+
"loss": 0.2375,
|
| 982 |
+
"step": 695
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.5384615384615384,
|
| 986 |
+
"grad_norm": 2.953125,
|
| 987 |
+
"learning_rate": 2.649870964949585e-05,
|
| 988 |
+
"loss": 0.2702,
|
| 989 |
+
"step": 700
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.5423076923076923,
|
| 993 |
+
"grad_norm": 1.4296875,
|
| 994 |
+
"learning_rate": 2.621290439379167e-05,
|
| 995 |
+
"loss": 0.2263,
|
| 996 |
+
"step": 705
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.5461538461538461,
|
| 1000 |
+
"grad_norm": 2.484375,
|
| 1001 |
+
"learning_rate": 2.5927307352501896e-05,
|
| 1002 |
+
"loss": 0.2527,
|
| 1003 |
+
"step": 710
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.55,
|
| 1007 |
+
"grad_norm": 2.296875,
|
| 1008 |
+
"learning_rate": 2.5641964726872526e-05,
|
| 1009 |
+
"loss": 0.1968,
|
| 1010 |
+
"step": 715
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.5538461538461539,
|
| 1014 |
+
"grad_norm": 2.015625,
|
| 1015 |
+
"learning_rate": 2.5356922676992524e-05,
|
| 1016 |
+
"loss": 0.2424,
|
| 1017 |
+
"step": 720
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.5576923076923077,
|
| 1021 |
+
"grad_norm": 1.9296875,
|
| 1022 |
+
"learning_rate": 2.5072227314326508e-05,
|
| 1023 |
+
"loss": 0.2826,
|
| 1024 |
+
"step": 725
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.5615384615384615,
|
| 1028 |
+
"grad_norm": 1.796875,
|
| 1029 |
+
"learning_rate": 2.4787924694255237e-05,
|
| 1030 |
+
"loss": 0.2383,
|
| 1031 |
+
"step": 730
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.5653846153846154,
|
| 1035 |
+
"grad_norm": 1.671875,
|
| 1036 |
+
"learning_rate": 2.4504060808625213e-05,
|
| 1037 |
+
"loss": 0.2624,
|
| 1038 |
+
"step": 735
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.5692307692307692,
|
| 1042 |
+
"grad_norm": 1.8125,
|
| 1043 |
+
"learning_rate": 2.422068157830854e-05,
|
| 1044 |
+
"loss": 0.2473,
|
| 1045 |
+
"step": 740
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.573076923076923,
|
| 1049 |
+
"grad_norm": 2.515625,
|
| 1050 |
+
"learning_rate": 2.3937832845774277e-05,
|
| 1051 |
+
"loss": 0.2391,
|
| 1052 |
+
"step": 745
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.5769230769230769,
|
| 1056 |
+
"grad_norm": 1.7265625,
|
| 1057 |
+
"learning_rate": 2.3655560367672454e-05,
|
| 1058 |
+
"loss": 0.251,
|
| 1059 |
+
"step": 750
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.5807692307692308,
|
| 1063 |
+
"grad_norm": 1.5390625,
|
| 1064 |
+
"learning_rate": 2.3373909807431993e-05,
|
| 1065 |
+
"loss": 0.2025,
|
| 1066 |
+
"step": 755
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.5846153846153846,
|
| 1070 |
+
"grad_norm": 1.6015625,
|
| 1071 |
+
"learning_rate": 2.3092926727873702e-05,
|
| 1072 |
+
"loss": 0.2399,
|
| 1073 |
+
"step": 760
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.5884615384615385,
|
| 1077 |
+
"grad_norm": 1.9375,
|
| 1078 |
+
"learning_rate": 2.2812656583839538e-05,
|
| 1079 |
+
"loss": 0.2689,
|
| 1080 |
+
"step": 765
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 0.5923076923076923,
|
| 1084 |
+
"grad_norm": 1.8515625,
|
| 1085 |
+
"learning_rate": 2.253314471483937e-05,
|
| 1086 |
+
"loss": 0.2576,
|
| 1087 |
+
"step": 770
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.5961538461538461,
|
| 1091 |
+
"grad_norm": 1.765625,
|
| 1092 |
+
"learning_rate": 2.2254436337716305e-05,
|
| 1093 |
+
"loss": 0.2349,
|
| 1094 |
+
"step": 775
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.6,
|
| 1098 |
+
"grad_norm": 1.421875,
|
| 1099 |
+
"learning_rate": 2.1976576539332024e-05,
|
| 1100 |
+
"loss": 0.2436,
|
| 1101 |
+
"step": 780
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.6038461538461538,
|
| 1105 |
+
"grad_norm": 1.890625,
|
| 1106 |
+
"learning_rate": 2.1699610269272952e-05,
|
| 1107 |
+
"loss": 0.2533,
|
| 1108 |
+
"step": 785
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 0.6076923076923076,
|
| 1112 |
+
"grad_norm": 1.515625,
|
| 1113 |
+
"learning_rate": 2.1423582332578796e-05,
|
| 1114 |
+
"loss": 0.2787,
|
| 1115 |
+
"step": 790
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.6115384615384616,
|
| 1119 |
+
"grad_norm": 1.703125,
|
| 1120 |
+
"learning_rate": 2.1148537382494353e-05,
|
| 1121 |
+
"loss": 0.2291,
|
| 1122 |
+
"step": 795
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 0.6153846153846154,
|
| 1126 |
+
"grad_norm": 1.515625,
|
| 1127 |
+
"learning_rate": 2.0874519913245977e-05,
|
| 1128 |
+
"loss": 0.2548,
|
| 1129 |
+
"step": 800
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.6192307692307693,
|
| 1133 |
+
"grad_norm": 1.703125,
|
| 1134 |
+
"learning_rate": 2.060157425284366e-05,
|
| 1135 |
+
"loss": 0.2128,
|
| 1136 |
+
"step": 805
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.6230769230769231,
|
| 1140 |
+
"grad_norm": 1.8046875,
|
| 1141 |
+
"learning_rate": 2.0329744555910075e-05,
|
| 1142 |
+
"loss": 0.2442,
|
| 1143 |
+
"step": 810
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.6269230769230769,
|
| 1147 |
+
"grad_norm": 1.5703125,
|
| 1148 |
+
"learning_rate": 2.0059074796537662e-05,
|
| 1149 |
+
"loss": 0.2584,
|
| 1150 |
+
"step": 815
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 0.6307692307692307,
|
| 1154 |
+
"grad_norm": 1.9375,
|
| 1155 |
+
"learning_rate": 1.978960876117491e-05,
|
| 1156 |
+
"loss": 0.2312,
|
| 1157 |
+
"step": 820
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.6346153846153846,
|
| 1161 |
+
"grad_norm": 2.59375,
|
| 1162 |
+
"learning_rate": 1.9521390041542946e-05,
|
| 1163 |
+
"loss": 0.2475,
|
| 1164 |
+
"step": 825
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 0.6384615384615384,
|
| 1168 |
+
"grad_norm": 2.203125,
|
| 1169 |
+
"learning_rate": 1.9254462027583753e-05,
|
| 1170 |
+
"loss": 0.2416,
|
| 1171 |
+
"step": 830
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.6423076923076924,
|
| 1175 |
+
"grad_norm": 1.578125,
|
| 1176 |
+
"learning_rate": 1.8988867900440845e-05,
|
| 1177 |
+
"loss": 0.2377,
|
| 1178 |
+
"step": 835
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.6461538461538462,
|
| 1182 |
+
"grad_norm": 1.796875,
|
| 1183 |
+
"learning_rate": 1.87246506254739e-05,
|
| 1184 |
+
"loss": 0.2644,
|
| 1185 |
+
"step": 840
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.65,
|
| 1189 |
+
"grad_norm": 1.71875,
|
| 1190 |
+
"learning_rate": 1.8461852945308196e-05,
|
| 1191 |
+
"loss": 0.2375,
|
| 1192 |
+
"step": 845
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 0.6538461538461539,
|
| 1196 |
+
"grad_norm": 1.875,
|
| 1197 |
+
"learning_rate": 1.8200517372920078e-05,
|
| 1198 |
+
"loss": 0.2666,
|
| 1199 |
+
"step": 850
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.6576923076923077,
|
| 1203 |
+
"grad_norm": 1.8671875,
|
| 1204 |
+
"learning_rate": 1.794068618475967e-05,
|
| 1205 |
+
"loss": 0.2515,
|
| 1206 |
+
"step": 855
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 0.6615384615384615,
|
| 1210 |
+
"grad_norm": 1.5703125,
|
| 1211 |
+
"learning_rate": 1.7682401413911697e-05,
|
| 1212 |
+
"loss": 0.2091,
|
| 1213 |
+
"step": 860
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.6653846153846154,
|
| 1217 |
+
"grad_norm": 1.8203125,
|
| 1218 |
+
"learning_rate": 1.742570484329583e-05,
|
| 1219 |
+
"loss": 0.2332,
|
| 1220 |
+
"step": 865
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 0.6692307692307692,
|
| 1224 |
+
"grad_norm": 1.5859375,
|
| 1225 |
+
"learning_rate": 1.7170637998907397e-05,
|
| 1226 |
+
"loss": 0.2388,
|
| 1227 |
+
"step": 870
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.6730769230769231,
|
| 1231 |
+
"grad_norm": 1.8515625,
|
| 1232 |
+
"learning_rate": 1.6917242143099704e-05,
|
| 1233 |
+
"loss": 0.2565,
|
| 1234 |
+
"step": 875
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 0.676923076923077,
|
| 1238 |
+
"grad_norm": 1.984375,
|
| 1239 |
+
"learning_rate": 1.6665558267908986e-05,
|
| 1240 |
+
"loss": 0.2539,
|
| 1241 |
+
"step": 880
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.6807692307692308,
|
| 1245 |
+
"grad_norm": 1.8359375,
|
| 1246 |
+
"learning_rate": 1.64156270884231e-05,
|
| 1247 |
+
"loss": 0.2241,
|
| 1248 |
+
"step": 885
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 0.6846153846153846,
|
| 1252 |
+
"grad_norm": 1.625,
|
| 1253 |
+
"learning_rate": 1.616748903619504e-05,
|
| 1254 |
+
"loss": 0.2219,
|
| 1255 |
+
"step": 890
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.6884615384615385,
|
| 1259 |
+
"grad_norm": 1.375,
|
| 1260 |
+
"learning_rate": 1.5921184252702192e-05,
|
| 1261 |
+
"loss": 0.1955,
|
| 1262 |
+
"step": 895
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 0.6923076923076923,
|
| 1266 |
+
"grad_norm": 2.3125,
|
| 1267 |
+
"learning_rate": 1.5676752582852794e-05,
|
| 1268 |
+
"loss": 0.2408,
|
| 1269 |
+
"step": 900
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.6961538461538461,
|
| 1273 |
+
"grad_norm": 1.4609375,
|
| 1274 |
+
"learning_rate": 1.543423356854004e-05,
|
| 1275 |
+
"loss": 0.2302,
|
| 1276 |
+
"step": 905
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 0.7,
|
| 1280 |
+
"grad_norm": 1.1796875,
|
| 1281 |
+
"learning_rate": 1.5193666442245402e-05,
|
| 1282 |
+
"loss": 0.2266,
|
| 1283 |
+
"step": 910
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.7038461538461539,
|
| 1287 |
+
"grad_norm": 1.75,
|
| 1288 |
+
"learning_rate": 1.4955090120692006e-05,
|
| 1289 |
+
"loss": 0.2251,
|
| 1290 |
+
"step": 915
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 0.7076923076923077,
|
| 1294 |
+
"grad_norm": 1.359375,
|
| 1295 |
+
"learning_rate": 1.4718543198548998e-05,
|
| 1296 |
+
"loss": 0.2318,
|
| 1297 |
+
"step": 920
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 0.7115384615384616,
|
| 1301 |
+
"grad_norm": 1.6640625,
|
| 1302 |
+
"learning_rate": 1.4484063942188076e-05,
|
| 1303 |
+
"loss": 0.249,
|
| 1304 |
+
"step": 925
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 0.7153846153846154,
|
| 1308 |
+
"grad_norm": 2.109375,
|
| 1309 |
+
"learning_rate": 1.4251690283493105e-05,
|
| 1310 |
+
"loss": 0.2817,
|
| 1311 |
+
"step": 930
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 0.7192307692307692,
|
| 1315 |
+
"grad_norm": 2.609375,
|
| 1316 |
+
"learning_rate": 1.402145981372384e-05,
|
| 1317 |
+
"loss": 0.2437,
|
| 1318 |
+
"step": 935
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 0.7230769230769231,
|
| 1322 |
+
"grad_norm": 2.078125,
|
| 1323 |
+
"learning_rate": 1.3793409777434775e-05,
|
| 1324 |
+
"loss": 0.2358,
|
| 1325 |
+
"step": 940
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 0.7269230769230769,
|
| 1329 |
+
"grad_norm": 1.6328125,
|
| 1330 |
+
"learning_rate": 1.3567577066450066e-05,
|
| 1331 |
+
"loss": 0.2126,
|
| 1332 |
+
"step": 945
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 0.7307692307692307,
|
| 1336 |
+
"grad_norm": 1.2734375,
|
| 1337 |
+
"learning_rate": 1.3343998213895464e-05,
|
| 1338 |
+
"loss": 0.2248,
|
| 1339 |
+
"step": 950
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 0.7346153846153847,
|
| 1343 |
+
"grad_norm": 1.7421875,
|
| 1344 |
+
"learning_rate": 1.3122709388288423e-05,
|
| 1345 |
+
"loss": 0.2528,
|
| 1346 |
+
"step": 955
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 0.7384615384615385,
|
| 1350 |
+
"grad_norm": 1.703125,
|
| 1351 |
+
"learning_rate": 1.2903746387686994e-05,
|
| 1352 |
+
"loss": 0.2311,
|
| 1353 |
+
"step": 960
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 0.7423076923076923,
|
| 1357 |
+
"grad_norm": 1.734375,
|
| 1358 |
+
"learning_rate": 1.268714463389879e-05,
|
| 1359 |
+
"loss": 0.2829,
|
| 1360 |
+
"step": 965
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 0.7461538461538462,
|
| 1364 |
+
"grad_norm": 1.4140625,
|
| 1365 |
+
"learning_rate": 1.2472939166750766e-05,
|
| 1366 |
+
"loss": 0.2245,
|
| 1367 |
+
"step": 970
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 0.75,
|
| 1371 |
+
"grad_norm": 1.71875,
|
| 1372 |
+
"learning_rate": 1.2261164638420832e-05,
|
| 1373 |
+
"loss": 0.2457,
|
| 1374 |
+
"step": 975
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 0.7538461538461538,
|
| 1378 |
+
"grad_norm": 1.9765625,
|
| 1379 |
+
"learning_rate": 1.2051855307832074e-05,
|
| 1380 |
+
"loss": 0.2374,
|
| 1381 |
+
"step": 980
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 0.7576923076923077,
|
| 1385 |
+
"grad_norm": 1.7734375,
|
| 1386 |
+
"learning_rate": 1.184504503511073e-05,
|
| 1387 |
+
"loss": 0.2299,
|
| 1388 |
+
"step": 985
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 0.7615384615384615,
|
| 1392 |
+
"grad_norm": 1.8515625,
|
| 1393 |
+
"learning_rate": 1.1640767276108584e-05,
|
| 1394 |
+
"loss": 0.2576,
|
| 1395 |
+
"step": 990
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 0.7653846153846153,
|
| 1399 |
+
"grad_norm": 1.6640625,
|
| 1400 |
+
"learning_rate": 1.1439055076990782e-05,
|
| 1401 |
+
"loss": 0.2014,
|
| 1402 |
+
"step": 995
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 0.7692307692307693,
|
| 1406 |
+
"grad_norm": 1.4375,
|
| 1407 |
+
"learning_rate": 1.123994106888995e-05,
|
| 1408 |
+
"loss": 0.2618,
|
| 1409 |
+
"step": 1000
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 0.7730769230769231,
|
| 1413 |
+
"grad_norm": 1.7890625,
|
| 1414 |
+
"learning_rate": 1.1043457462627404e-05,
|
| 1415 |
+
"loss": 0.2112,
|
| 1416 |
+
"step": 1005
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 0.7769230769230769,
|
| 1420 |
+
"grad_norm": 1.5,
|
| 1421 |
+
"learning_rate": 1.084963604350241e-05,
|
| 1422 |
+
"loss": 0.236,
|
| 1423 |
+
"step": 1010
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 0.7807692307692308,
|
| 1427 |
+
"grad_norm": 1.328125,
|
| 1428 |
+
"learning_rate": 1.0658508166150224e-05,
|
| 1429 |
+
"loss": 0.221,
|
| 1430 |
+
"step": 1015
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 0.7846153846153846,
|
| 1434 |
+
"grad_norm": 2.15625,
|
| 1435 |
+
"learning_rate": 1.0470104749469867e-05,
|
| 1436 |
+
"loss": 0.2579,
|
| 1437 |
+
"step": 1020
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 0.7884615384615384,
|
| 1441 |
+
"grad_norm": 1.84375,
|
| 1442 |
+
"learning_rate": 1.0284456271622295e-05,
|
| 1443 |
+
"loss": 0.2139,
|
| 1444 |
+
"step": 1025
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 0.7923076923076923,
|
| 1448 |
+
"grad_norm": 1.8046875,
|
| 1449 |
+
"learning_rate": 1.010159276509999e-05,
|
| 1450 |
+
"loss": 0.2461,
|
| 1451 |
+
"step": 1030
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 0.7961538461538461,
|
| 1455 |
+
"grad_norm": 1.5625,
|
| 1456 |
+
"learning_rate": 9.921543811868548e-06,
|
| 1457 |
+
"loss": 0.2538,
|
| 1458 |
+
"step": 1035
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 0.8,
|
| 1462 |
+
"grad_norm": 1.65625,
|
| 1463 |
+
"learning_rate": 9.744338538581147e-06,
|
| 1464 |
+
"loss": 0.2173,
|
| 1465 |
+
"step": 1040
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 0.8038461538461539,
|
| 1469 |
+
"grad_norm": 1.4609375,
|
| 1470 |
+
"learning_rate": 9.570005611866787e-06,
|
| 1471 |
+
"loss": 0.2277,
|
| 1472 |
+
"step": 1045
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 0.8076923076923077,
|
| 1476 |
+
"grad_norm": 1.4609375,
|
| 1477 |
+
"learning_rate": 9.398573233692802e-06,
|
| 1478 |
+
"loss": 0.2432,
|
| 1479 |
+
"step": 1050
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 0.8115384615384615,
|
| 1483 |
+
"grad_norm": 1.4609375,
|
| 1484 |
+
"learning_rate": 9.230069136802633e-06,
|
| 1485 |
+
"loss": 0.2622,
|
| 1486 |
+
"step": 1055
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 0.8153846153846154,
|
| 1490 |
+
"grad_norm": 1.890625,
|
| 1491 |
+
"learning_rate": 9.064520580229482e-06,
|
| 1492 |
+
"loss": 0.2528,
|
| 1493 |
+
"step": 1060
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 0.8192307692307692,
|
| 1497 |
+
"grad_norm": 1.65625,
|
| 1498 |
+
"learning_rate": 8.90195434488659e-06,
|
| 1499 |
+
"loss": 0.2785,
|
| 1500 |
+
"step": 1065
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 0.823076923076923,
|
| 1504 |
+
"grad_norm": 1.6484375,
|
| 1505 |
+
"learning_rate": 8.742396729234873e-06,
|
| 1506 |
+
"loss": 0.2489,
|
| 1507 |
+
"step": 1070
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 0.8269230769230769,
|
| 1511 |
+
"grad_norm": 1.71875,
|
| 1512 |
+
"learning_rate": 8.5858735450286e-06,
|
| 1513 |
+
"loss": 0.2395,
|
| 1514 |
+
"step": 1075
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 0.8307692307692308,
|
| 1518 |
+
"grad_norm": 2.578125,
|
| 1519 |
+
"learning_rate": 8.432410113139803e-06,
|
| 1520 |
+
"loss": 0.2466,
|
| 1521 |
+
"step": 1080
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 0.8346153846153846,
|
| 1525 |
+
"grad_norm": 2.0625,
|
| 1526 |
+
"learning_rate": 8.2820312594621e-06,
|
| 1527 |
+
"loss": 0.261,
|
| 1528 |
+
"step": 1085
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"epoch": 0.8384615384615385,
|
| 1532 |
+
"grad_norm": 1.6875,
|
| 1533 |
+
"learning_rate": 8.134761310894623e-06,
|
| 1534 |
+
"loss": 0.2618,
|
| 1535 |
+
"step": 1090
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 0.8423076923076923,
|
| 1539 |
+
"grad_norm": 1.7109375,
|
| 1540 |
+
"learning_rate": 7.990624091406568e-06,
|
| 1541 |
+
"loss": 0.2358,
|
| 1542 |
+
"step": 1095
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 0.8461538461538461,
|
| 1546 |
+
"grad_norm": 1.40625,
|
| 1547 |
+
"learning_rate": 7.849642918183244e-06,
|
| 1548 |
+
"loss": 0.2252,
|
| 1549 |
+
"step": 1100
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 0.85,
|
| 1553 |
+
"grad_norm": 2.640625,
|
| 1554 |
+
"learning_rate": 7.711840597853998e-06,
|
| 1555 |
+
"loss": 0.25,
|
| 1556 |
+
"step": 1105
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 0.8538461538461538,
|
| 1560 |
+
"grad_norm": 1.546875,
|
| 1561 |
+
"learning_rate": 7.57723942280278e-06,
|
| 1562 |
+
"loss": 0.2206,
|
| 1563 |
+
"step": 1110
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 0.8576923076923076,
|
| 1567 |
+
"grad_norm": 1.3359375,
|
| 1568 |
+
"learning_rate": 7.44586116756188e-06,
|
| 1569 |
+
"loss": 0.2147,
|
| 1570 |
+
"step": 1115
|
| 1571 |
+
},
|
| 1572 |
+
{
|
| 1573 |
+
"epoch": 0.8615384615384616,
|
| 1574 |
+
"grad_norm": 1.3671875,
|
| 1575 |
+
"learning_rate": 7.317727085289458e-06,
|
| 1576 |
+
"loss": 0.2299,
|
| 1577 |
+
"step": 1120
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 0.8653846153846154,
|
| 1581 |
+
"grad_norm": 1.4453125,
|
| 1582 |
+
"learning_rate": 7.192857904331393e-06,
|
| 1583 |
+
"loss": 0.2149,
|
| 1584 |
+
"step": 1125
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 0.8692307692307693,
|
| 1588 |
+
"grad_norm": 1.609375,
|
| 1589 |
+
"learning_rate": 7.071273824868049e-06,
|
| 1590 |
+
"loss": 0.2584,
|
| 1591 |
+
"step": 1130
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 0.8730769230769231,
|
| 1595 |
+
"grad_norm": 1.1953125,
|
| 1596 |
+
"learning_rate": 6.95299451564648e-06,
|
| 1597 |
+
"loss": 0.2256,
|
| 1598 |
+
"step": 1135
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 0.8769230769230769,
|
| 1602 |
+
"grad_norm": 1.6484375,
|
| 1603 |
+
"learning_rate": 6.838039110798579e-06,
|
| 1604 |
+
"loss": 0.2801,
|
| 1605 |
+
"step": 1140
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 0.8807692307692307,
|
| 1609 |
+
"grad_norm": 1.796875,
|
| 1610 |
+
"learning_rate": 6.726426206745781e-06,
|
| 1611 |
+
"loss": 0.2297,
|
| 1612 |
+
"step": 1145
|
| 1613 |
+
},
|
| 1614 |
+
{
|
| 1615 |
+
"epoch": 0.8846153846153846,
|
| 1616 |
+
"grad_norm": 1.328125,
|
| 1617 |
+
"learning_rate": 6.6181738591906685e-06,
|
| 1618 |
+
"loss": 0.2127,
|
| 1619 |
+
"step": 1150
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 0.8884615384615384,
|
| 1623 |
+
"grad_norm": 1.8515625,
|
| 1624 |
+
"learning_rate": 6.513299580196086e-06,
|
| 1625 |
+
"loss": 0.2394,
|
| 1626 |
+
"step": 1155
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 0.8923076923076924,
|
| 1630 |
+
"grad_norm": 1.5234375,
|
| 1631 |
+
"learning_rate": 6.411820335352222e-06,
|
| 1632 |
+
"loss": 0.2193,
|
| 1633 |
+
"step": 1160
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 0.8961538461538462,
|
| 1637 |
+
"grad_norm": 1.78125,
|
| 1638 |
+
"learning_rate": 6.313752541032071e-06,
|
| 1639 |
+
"loss": 0.233,
|
| 1640 |
+
"step": 1165
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 0.9,
|
| 1644 |
+
"grad_norm": 2.296875,
|
| 1645 |
+
"learning_rate": 6.219112061735721e-06,
|
| 1646 |
+
"loss": 0.2243,
|
| 1647 |
+
"step": 1170
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 0.9038461538461539,
|
| 1651 |
+
"grad_norm": 2.0625,
|
| 1652 |
+
"learning_rate": 6.127914207523965e-06,
|
| 1653 |
+
"loss": 0.231,
|
| 1654 |
+
"step": 1175
|
| 1655 |
+
},
|
| 1656 |
+
{
|
| 1657 |
+
"epoch": 0.9076923076923077,
|
| 1658 |
+
"grad_norm": 1.2421875,
|
| 1659 |
+
"learning_rate": 6.040173731541575e-06,
|
| 1660 |
+
"loss": 0.1955,
|
| 1661 |
+
"step": 1180
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 0.9115384615384615,
|
| 1665 |
+
"grad_norm": 1.6328125,
|
| 1666 |
+
"learning_rate": 5.955904827630666e-06,
|
| 1667 |
+
"loss": 0.2635,
|
| 1668 |
+
"step": 1185
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 0.9153846153846154,
|
| 1672 |
+
"grad_norm": 1.5625,
|
| 1673 |
+
"learning_rate": 5.87512112803456e-06,
|
| 1674 |
+
"loss": 0.2445,
|
| 1675 |
+
"step": 1190
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 0.9192307692307692,
|
| 1679 |
+
"grad_norm": 2.59375,
|
| 1680 |
+
"learning_rate": 5.797835701192484e-06,
|
| 1681 |
+
"loss": 0.23,
|
| 1682 |
+
"step": 1195
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 0.9230769230769231,
|
| 1686 |
+
"grad_norm": 1.7109375,
|
| 1687 |
+
"learning_rate": 5.7240610496254724e-06,
|
| 1688 |
+
"loss": 0.2614,
|
| 1689 |
+
"step": 1200
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 0.926923076923077,
|
| 1693 |
+
"grad_norm": 1.65625,
|
| 1694 |
+
"learning_rate": 5.65380910791383e-06,
|
| 1695 |
+
"loss": 0.251,
|
| 1696 |
+
"step": 1205
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 0.9307692307692308,
|
| 1700 |
+
"grad_norm": 1.3046875,
|
| 1701 |
+
"learning_rate": 5.5870912407664715e-06,
|
| 1702 |
+
"loss": 0.2453,
|
| 1703 |
+
"step": 1210
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 0.9346153846153846,
|
| 1707 |
+
"grad_norm": 1.9453125,
|
| 1708 |
+
"learning_rate": 5.52391824118243e-06,
|
| 1709 |
+
"loss": 0.2123,
|
| 1710 |
+
"step": 1215
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 0.9384615384615385,
|
| 1714 |
+
"grad_norm": 1.4296875,
|
| 1715 |
+
"learning_rate": 5.464300328704876e-06,
|
| 1716 |
+
"loss": 0.2387,
|
| 1717 |
+
"step": 1220
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 0.9423076923076923,
|
| 1721 |
+
"grad_norm": 1.625,
|
| 1722 |
+
"learning_rate": 5.408247147767904e-06,
|
| 1723 |
+
"loss": 0.2374,
|
| 1724 |
+
"step": 1225
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 0.9461538461538461,
|
| 1728 |
+
"grad_norm": 1.9765625,
|
| 1729 |
+
"learning_rate": 5.355767766136315e-06,
|
| 1730 |
+
"loss": 0.229,
|
| 1731 |
+
"step": 1230
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 0.95,
|
| 1735 |
+
"grad_norm": 1.75,
|
| 1736 |
+
"learning_rate": 5.3068706734387484e-06,
|
| 1737 |
+
"loss": 0.2112,
|
| 1738 |
+
"step": 1235
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 0.9538461538461539,
|
| 1742 |
+
"grad_norm": 1.5234375,
|
| 1743 |
+
"learning_rate": 5.261563779794289e-06,
|
| 1744 |
+
"loss": 0.2267,
|
| 1745 |
+
"step": 1240
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 0.9576923076923077,
|
| 1749 |
+
"grad_norm": 1.71875,
|
| 1750 |
+
"learning_rate": 5.219854414532854e-06,
|
| 1751 |
+
"loss": 0.2318,
|
| 1752 |
+
"step": 1245
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 0.9615384615384616,
|
| 1756 |
+
"grad_norm": 2.234375,
|
| 1757 |
+
"learning_rate": 5.181749325009511e-06,
|
| 1758 |
+
"loss": 0.2305,
|
| 1759 |
+
"step": 1250
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 0.9653846153846154,
|
| 1763 |
+
"grad_norm": 1.8359375,
|
| 1764 |
+
"learning_rate": 5.1472546755129605e-06,
|
| 1765 |
+
"loss": 0.275,
|
| 1766 |
+
"step": 1255
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 0.9692307692307692,
|
| 1770 |
+
"grad_norm": 1.5390625,
|
| 1771 |
+
"learning_rate": 5.116376046268333e-06,
|
| 1772 |
+
"loss": 0.2549,
|
| 1773 |
+
"step": 1260
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 0.9730769230769231,
|
| 1777 |
+
"grad_norm": 1.3984375,
|
| 1778 |
+
"learning_rate": 5.089118432534466e-06,
|
| 1779 |
+
"loss": 0.2493,
|
| 1780 |
+
"step": 1265
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 0.9769230769230769,
|
| 1784 |
+
"grad_norm": 1.8671875,
|
| 1785 |
+
"learning_rate": 5.065486243795829e-06,
|
| 1786 |
+
"loss": 0.263,
|
| 1787 |
+
"step": 1270
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 0.9807692307692307,
|
| 1791 |
+
"grad_norm": 1.5390625,
|
| 1792 |
+
"learning_rate": 5.045483303049184e-06,
|
| 1793 |
+
"loss": 0.2837,
|
| 1794 |
+
"step": 1275
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 0.9846153846153847,
|
| 1798 |
+
"grad_norm": 1.296875,
|
| 1799 |
+
"learning_rate": 5.02911284618515e-06,
|
| 1800 |
+
"loss": 0.1995,
|
| 1801 |
+
"step": 1280
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 0.9884615384615385,
|
| 1805 |
+
"grad_norm": 1.4375,
|
| 1806 |
+
"learning_rate": 5.01637752146472e-06,
|
| 1807 |
+
"loss": 0.2232,
|
| 1808 |
+
"step": 1285
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 0.9923076923076923,
|
| 1812 |
+
"grad_norm": 2.015625,
|
| 1813 |
+
"learning_rate": 5.007279389090856e-06,
|
| 1814 |
+
"loss": 0.2487,
|
| 1815 |
+
"step": 1290
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 0.9961538461538462,
|
| 1819 |
+
"grad_norm": 1.84375,
|
| 1820 |
+
"learning_rate": 5.001819920875201e-06,
|
| 1821 |
+
"loss": 0.2508,
|
| 1822 |
+
"step": 1295
|
| 1823 |
+
},
|
| 1824 |
+
{
|
| 1825 |
+
"epoch": 1.0,
|
| 1826 |
+
"grad_norm": 1.8359375,
|
| 1827 |
+
"learning_rate": 5e-06,
|
| 1828 |
+
"loss": 0.2403,
|
| 1829 |
+
"step": 1300
|
| 1830 |
+
},
|
| 1831 |
{
|
| 1832 |
"epoch": 1.0,
|
| 1833 |
+
"step": 1300,
|
| 1834 |
+
"total_flos": 4.404961752383488e+16,
|
| 1835 |
+
"train_loss": 0.28405343046555154,
|
| 1836 |
+
"train_runtime": 1315.0218,
|
| 1837 |
+
"train_samples_per_second": 0.989,
|
| 1838 |
+
"train_steps_per_second": 0.989
|
| 1839 |
}
|
| 1840 |
],
|
| 1841 |
"logging_steps": 5,
|
| 1842 |
+
"max_steps": 1300,
|
| 1843 |
"num_input_tokens_seen": 0,
|
| 1844 |
"num_train_epochs": 1,
|
| 1845 |
"save_steps": 100,
|
|
|
|
| 1855 |
"attributes": {}
|
| 1856 |
}
|
| 1857 |
},
|
| 1858 |
+
"total_flos": 4.404961752383488e+16,
|
| 1859 |
"train_batch_size": 1,
|
| 1860 |
"trial_name": null,
|
| 1861 |
"trial_params": null
|