End of training
Browse files- README.md +2 -1
- all_results.json +8 -0
- train_results.json +8 -0
- trainer_state.json +3151 -0
- training_loss.png +0 -0
README.md
CHANGED
|
@@ -4,6 +4,7 @@ license: apache-2.0
|
|
| 4 |
base_model: Qwen/Qwen3-1.7B
|
| 5 |
tags:
|
| 6 |
- llama-factory
|
|
|
|
| 7 |
- generated_from_trainer
|
| 8 |
model-index:
|
| 9 |
- name: Qwen3-1.7B-SFT-TinyV_Reasoning_Balanced_v2.1_Qwen3-LR1.0e-5-EPOCHS2
|
|
@@ -15,7 +16,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 15 |
|
| 16 |
# Qwen3-1.7B-SFT-TinyV_Reasoning_Balanced_v2.1_Qwen3-LR1.0e-5-EPOCHS2
|
| 17 |
|
| 18 |
-
This model is a fine-tuned version of [Qwen/Qwen3-1.7B](https://huggingface.co/Qwen/Qwen3-1.7B) on
|
| 19 |
|
| 20 |
## Model description
|
| 21 |
|
|
|
|
| 4 |
base_model: Qwen/Qwen3-1.7B
|
| 5 |
tags:
|
| 6 |
- llama-factory
|
| 7 |
+
- full
|
| 8 |
- generated_from_trainer
|
| 9 |
model-index:
|
| 10 |
- name: Qwen3-1.7B-SFT-TinyV_Reasoning_Balanced_v2.1_Qwen3-LR1.0e-5-EPOCHS2
|
|
|
|
| 16 |
|
| 17 |
# Qwen3-1.7B-SFT-TinyV_Reasoning_Balanced_v2.1_Qwen3-LR1.0e-5-EPOCHS2
|
| 18 |
|
| 19 |
+
This model is a fine-tuned version of [Qwen/Qwen3-1.7B](https://huggingface.co/Qwen/Qwen3-1.7B) on the TinyV_Reasoning_Balanced_v2.1_Qwen3 dataset.
|
| 20 |
|
| 21 |
## Model description
|
| 22 |
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.0,
|
| 3 |
+
"total_flos": 6.347064608565494e+17,
|
| 4 |
+
"train_loss": 0.30420903603101634,
|
| 5 |
+
"train_runtime": 1521.5027,
|
| 6 |
+
"train_samples_per_second": 148.979,
|
| 7 |
+
"train_steps_per_second": 0.292
|
| 8 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.0,
|
| 3 |
+
"total_flos": 6.347064608565494e+17,
|
| 4 |
+
"train_loss": 0.30420903603101634,
|
| 5 |
+
"train_runtime": 1521.5027,
|
| 6 |
+
"train_samples_per_second": 148.979,
|
| 7 |
+
"train_steps_per_second": 0.292
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3151 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 2.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 444,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.004517221908526256,
|
| 14 |
+
"grad_norm": 120.71863555908203,
|
| 15 |
+
"learning_rate": 0.0,
|
| 16 |
+
"loss": 1.9359,
|
| 17 |
+
"step": 1
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.009034443817052512,
|
| 21 |
+
"grad_norm": 116.67710876464844,
|
| 22 |
+
"learning_rate": 2.2222222222222224e-07,
|
| 23 |
+
"loss": 1.9035,
|
| 24 |
+
"step": 2
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.01355166572557877,
|
| 28 |
+
"grad_norm": 116.16535949707031,
|
| 29 |
+
"learning_rate": 4.444444444444445e-07,
|
| 30 |
+
"loss": 1.872,
|
| 31 |
+
"step": 3
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.018068887634105024,
|
| 35 |
+
"grad_norm": 115.42186737060547,
|
| 36 |
+
"learning_rate": 6.666666666666667e-07,
|
| 37 |
+
"loss": 1.8759,
|
| 38 |
+
"step": 4
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.02258610954263128,
|
| 42 |
+
"grad_norm": 113.68812561035156,
|
| 43 |
+
"learning_rate": 8.88888888888889e-07,
|
| 44 |
+
"loss": 1.8433,
|
| 45 |
+
"step": 5
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.02710333145115754,
|
| 49 |
+
"grad_norm": 108.90851593017578,
|
| 50 |
+
"learning_rate": 1.111111111111111e-06,
|
| 51 |
+
"loss": 1.8224,
|
| 52 |
+
"step": 6
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.03162055335968379,
|
| 56 |
+
"grad_norm": 105.59119415283203,
|
| 57 |
+
"learning_rate": 1.3333333333333334e-06,
|
| 58 |
+
"loss": 1.7881,
|
| 59 |
+
"step": 7
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.03613777526821005,
|
| 63 |
+
"grad_norm": 78.03402709960938,
|
| 64 |
+
"learning_rate": 1.5555555555555558e-06,
|
| 65 |
+
"loss": 1.5911,
|
| 66 |
+
"step": 8
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 0.040654997176736304,
|
| 70 |
+
"grad_norm": 71.16693115234375,
|
| 71 |
+
"learning_rate": 1.777777777777778e-06,
|
| 72 |
+
"loss": 1.5803,
|
| 73 |
+
"step": 9
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.04517221908526256,
|
| 77 |
+
"grad_norm": 37.84841537475586,
|
| 78 |
+
"learning_rate": 2.0000000000000003e-06,
|
| 79 |
+
"loss": 1.3045,
|
| 80 |
+
"step": 10
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.049689440993788817,
|
| 84 |
+
"grad_norm": 36.52073669433594,
|
| 85 |
+
"learning_rate": 2.222222222222222e-06,
|
| 86 |
+
"loss": 1.2981,
|
| 87 |
+
"step": 11
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.05420666290231508,
|
| 91 |
+
"grad_norm": 32.27294158935547,
|
| 92 |
+
"learning_rate": 2.4444444444444447e-06,
|
| 93 |
+
"loss": 1.2347,
|
| 94 |
+
"step": 12
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.058723884810841336,
|
| 98 |
+
"grad_norm": 30.345687866210938,
|
| 99 |
+
"learning_rate": 2.666666666666667e-06,
|
| 100 |
+
"loss": 1.2005,
|
| 101 |
+
"step": 13
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 0.06324110671936758,
|
| 105 |
+
"grad_norm": 18.027023315429688,
|
| 106 |
+
"learning_rate": 2.888888888888889e-06,
|
| 107 |
+
"loss": 0.8885,
|
| 108 |
+
"step": 14
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 0.06775832862789384,
|
| 112 |
+
"grad_norm": 17.39182472229004,
|
| 113 |
+
"learning_rate": 3.1111111111111116e-06,
|
| 114 |
+
"loss": 0.8686,
|
| 115 |
+
"step": 15
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.0722755505364201,
|
| 119 |
+
"grad_norm": 10.756152153015137,
|
| 120 |
+
"learning_rate": 3.3333333333333333e-06,
|
| 121 |
+
"loss": 0.8102,
|
| 122 |
+
"step": 16
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.07679277244494635,
|
| 126 |
+
"grad_norm": 8.62587833404541,
|
| 127 |
+
"learning_rate": 3.555555555555556e-06,
|
| 128 |
+
"loss": 0.7657,
|
| 129 |
+
"step": 17
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.08130999435347261,
|
| 133 |
+
"grad_norm": 5.619058609008789,
|
| 134 |
+
"learning_rate": 3.777777777777778e-06,
|
| 135 |
+
"loss": 0.7018,
|
| 136 |
+
"step": 18
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 0.08582721626199886,
|
| 140 |
+
"grad_norm": 5.939872741699219,
|
| 141 |
+
"learning_rate": 4.000000000000001e-06,
|
| 142 |
+
"loss": 0.7027,
|
| 143 |
+
"step": 19
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 0.09034443817052512,
|
| 147 |
+
"grad_norm": 5.757226943969727,
|
| 148 |
+
"learning_rate": 4.222222222222223e-06,
|
| 149 |
+
"loss": 0.6736,
|
| 150 |
+
"step": 20
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 0.09486166007905138,
|
| 154 |
+
"grad_norm": 5.253839492797852,
|
| 155 |
+
"learning_rate": 4.444444444444444e-06,
|
| 156 |
+
"loss": 0.6412,
|
| 157 |
+
"step": 21
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 0.09937888198757763,
|
| 161 |
+
"grad_norm": 4.535945415496826,
|
| 162 |
+
"learning_rate": 4.666666666666667e-06,
|
| 163 |
+
"loss": 0.6275,
|
| 164 |
+
"step": 22
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.1038961038961039,
|
| 168 |
+
"grad_norm": 3.7249350547790527,
|
| 169 |
+
"learning_rate": 4.888888888888889e-06,
|
| 170 |
+
"loss": 0.6058,
|
| 171 |
+
"step": 23
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.10841332580463016,
|
| 175 |
+
"grad_norm": 2.893019914627075,
|
| 176 |
+
"learning_rate": 5.1111111111111115e-06,
|
| 177 |
+
"loss": 0.5847,
|
| 178 |
+
"step": 24
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 0.11293054771315642,
|
| 182 |
+
"grad_norm": 2.4801812171936035,
|
| 183 |
+
"learning_rate": 5.333333333333334e-06,
|
| 184 |
+
"loss": 0.5557,
|
| 185 |
+
"step": 25
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.11744776962168267,
|
| 189 |
+
"grad_norm": 2.7257401943206787,
|
| 190 |
+
"learning_rate": 5.555555555555557e-06,
|
| 191 |
+
"loss": 0.5552,
|
| 192 |
+
"step": 26
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 0.12196499153020893,
|
| 196 |
+
"grad_norm": 2.7159502506256104,
|
| 197 |
+
"learning_rate": 5.777777777777778e-06,
|
| 198 |
+
"loss": 0.5313,
|
| 199 |
+
"step": 27
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 0.12648221343873517,
|
| 203 |
+
"grad_norm": 2.2152352333068848,
|
| 204 |
+
"learning_rate": 6e-06,
|
| 205 |
+
"loss": 0.4959,
|
| 206 |
+
"step": 28
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.13099943534726144,
|
| 210 |
+
"grad_norm": 1.7073229551315308,
|
| 211 |
+
"learning_rate": 6.222222222222223e-06,
|
| 212 |
+
"loss": 0.4862,
|
| 213 |
+
"step": 29
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.13551665725578768,
|
| 217 |
+
"grad_norm": 1.622607707977295,
|
| 218 |
+
"learning_rate": 6.444444444444445e-06,
|
| 219 |
+
"loss": 0.4671,
|
| 220 |
+
"step": 30
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 0.14003387916431395,
|
| 224 |
+
"grad_norm": 1.7897526025772095,
|
| 225 |
+
"learning_rate": 6.666666666666667e-06,
|
| 226 |
+
"loss": 0.4829,
|
| 227 |
+
"step": 31
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 0.1445511010728402,
|
| 231 |
+
"grad_norm": 1.609177827835083,
|
| 232 |
+
"learning_rate": 6.88888888888889e-06,
|
| 233 |
+
"loss": 0.4521,
|
| 234 |
+
"step": 32
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.14906832298136646,
|
| 238 |
+
"grad_norm": 1.6080033779144287,
|
| 239 |
+
"learning_rate": 7.111111111111112e-06,
|
| 240 |
+
"loss": 0.4521,
|
| 241 |
+
"step": 33
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 0.1535855448898927,
|
| 245 |
+
"grad_norm": 1.6252771615982056,
|
| 246 |
+
"learning_rate": 7.333333333333333e-06,
|
| 247 |
+
"loss": 0.4435,
|
| 248 |
+
"step": 34
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.15810276679841898,
|
| 252 |
+
"grad_norm": 1.5607670545578003,
|
| 253 |
+
"learning_rate": 7.555555555555556e-06,
|
| 254 |
+
"loss": 0.4279,
|
| 255 |
+
"step": 35
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.16261998870694522,
|
| 259 |
+
"grad_norm": 1.6145962476730347,
|
| 260 |
+
"learning_rate": 7.77777777777778e-06,
|
| 261 |
+
"loss": 0.4181,
|
| 262 |
+
"step": 36
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 0.1671372106154715,
|
| 266 |
+
"grad_norm": 1.5099209547042847,
|
| 267 |
+
"learning_rate": 8.000000000000001e-06,
|
| 268 |
+
"loss": 0.4247,
|
| 269 |
+
"step": 37
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 0.17165443252399773,
|
| 273 |
+
"grad_norm": 1.5651538372039795,
|
| 274 |
+
"learning_rate": 8.222222222222222e-06,
|
| 275 |
+
"loss": 0.4123,
|
| 276 |
+
"step": 38
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 0.176171654432524,
|
| 280 |
+
"grad_norm": 1.4987351894378662,
|
| 281 |
+
"learning_rate": 8.444444444444446e-06,
|
| 282 |
+
"loss": 0.395,
|
| 283 |
+
"step": 39
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 0.18068887634105024,
|
| 287 |
+
"grad_norm": 1.3249200582504272,
|
| 288 |
+
"learning_rate": 8.666666666666668e-06,
|
| 289 |
+
"loss": 0.3852,
|
| 290 |
+
"step": 40
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.1852060982495765,
|
| 294 |
+
"grad_norm": 1.407200813293457,
|
| 295 |
+
"learning_rate": 8.888888888888888e-06,
|
| 296 |
+
"loss": 0.3914,
|
| 297 |
+
"step": 41
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.18972332015810275,
|
| 301 |
+
"grad_norm": 1.3914618492126465,
|
| 302 |
+
"learning_rate": 9.111111111111112e-06,
|
| 303 |
+
"loss": 0.3722,
|
| 304 |
+
"step": 42
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 0.19424054206662902,
|
| 308 |
+
"grad_norm": 1.36204195022583,
|
| 309 |
+
"learning_rate": 9.333333333333334e-06,
|
| 310 |
+
"loss": 0.3632,
|
| 311 |
+
"step": 43
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 0.19875776397515527,
|
| 315 |
+
"grad_norm": 1.3414491415023804,
|
| 316 |
+
"learning_rate": 9.555555555555556e-06,
|
| 317 |
+
"loss": 0.3827,
|
| 318 |
+
"step": 44
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 0.20327498588368154,
|
| 322 |
+
"grad_norm": 1.4723490476608276,
|
| 323 |
+
"learning_rate": 9.777777777777779e-06,
|
| 324 |
+
"loss": 0.3779,
|
| 325 |
+
"step": 45
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.2077922077922078,
|
| 329 |
+
"grad_norm": 1.2399615049362183,
|
| 330 |
+
"learning_rate": 1e-05,
|
| 331 |
+
"loss": 0.3632,
|
| 332 |
+
"step": 46
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.21230942970073405,
|
| 336 |
+
"grad_norm": 1.4166227579116821,
|
| 337 |
+
"learning_rate": 9.999845014267928e-06,
|
| 338 |
+
"loss": 0.3567,
|
| 339 |
+
"step": 47
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.21682665160926032,
|
| 343 |
+
"grad_norm": 1.5010066032409668,
|
| 344 |
+
"learning_rate": 9.999380066679943e-06,
|
| 345 |
+
"loss": 0.3489,
|
| 346 |
+
"step": 48
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 0.22134387351778656,
|
| 350 |
+
"grad_norm": 1.3002145290374756,
|
| 351 |
+
"learning_rate": 9.998605186060138e-06,
|
| 352 |
+
"loss": 0.3631,
|
| 353 |
+
"step": 49
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 0.22586109542631283,
|
| 357 |
+
"grad_norm": 1.2854527235031128,
|
| 358 |
+
"learning_rate": 9.997520420446694e-06,
|
| 359 |
+
"loss": 0.3407,
|
| 360 |
+
"step": 50
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 0.23037831733483907,
|
| 364 |
+
"grad_norm": 1.210677981376648,
|
| 365 |
+
"learning_rate": 9.996125837088883e-06,
|
| 366 |
+
"loss": 0.356,
|
| 367 |
+
"step": 51
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 0.23489553924336534,
|
| 371 |
+
"grad_norm": 1.233259916305542,
|
| 372 |
+
"learning_rate": 9.99442152244292e-06,
|
| 373 |
+
"loss": 0.3356,
|
| 374 |
+
"step": 52
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.23941276115189158,
|
| 378 |
+
"grad_norm": 1.253690481185913,
|
| 379 |
+
"learning_rate": 9.992407582166582e-06,
|
| 380 |
+
"loss": 0.3518,
|
| 381 |
+
"step": 53
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 0.24392998306041785,
|
| 385 |
+
"grad_norm": 1.2361208200454712,
|
| 386 |
+
"learning_rate": 9.990084141112674e-06,
|
| 387 |
+
"loss": 0.3272,
|
| 388 |
+
"step": 54
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 0.2484472049689441,
|
| 392 |
+
"grad_norm": 1.2044817209243774,
|
| 393 |
+
"learning_rate": 9.98745134332128e-06,
|
| 394 |
+
"loss": 0.3391,
|
| 395 |
+
"step": 55
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 0.25296442687747034,
|
| 399 |
+
"grad_norm": 1.3248499631881714,
|
| 400 |
+
"learning_rate": 9.984509352010839e-06,
|
| 401 |
+
"loss": 0.3439,
|
| 402 |
+
"step": 56
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 0.2574816487859966,
|
| 406 |
+
"grad_norm": 1.2829524278640747,
|
| 407 |
+
"learning_rate": 9.981258349568018e-06,
|
| 408 |
+
"loss": 0.3248,
|
| 409 |
+
"step": 57
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 0.2619988706945229,
|
| 413 |
+
"grad_norm": 1.1910191774368286,
|
| 414 |
+
"learning_rate": 9.97769853753642e-06,
|
| 415 |
+
"loss": 0.337,
|
| 416 |
+
"step": 58
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.26651609260304915,
|
| 420 |
+
"grad_norm": 1.0255918502807617,
|
| 421 |
+
"learning_rate": 9.973830136604068e-06,
|
| 422 |
+
"loss": 0.3275,
|
| 423 |
+
"step": 59
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 0.27103331451157536,
|
| 427 |
+
"grad_norm": 1.2948644161224365,
|
| 428 |
+
"learning_rate": 9.969653386589749e-06,
|
| 429 |
+
"loss": 0.3329,
|
| 430 |
+
"step": 60
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 0.27555053642010163,
|
| 434 |
+
"grad_norm": 1.26154625415802,
|
| 435 |
+
"learning_rate": 9.965168546428122e-06,
|
| 436 |
+
"loss": 0.3306,
|
| 437 |
+
"step": 61
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 0.2800677583286279,
|
| 441 |
+
"grad_norm": 1.1000076532363892,
|
| 442 |
+
"learning_rate": 9.960375894153682e-06,
|
| 443 |
+
"loss": 0.3208,
|
| 444 |
+
"step": 62
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 0.2845849802371542,
|
| 448 |
+
"grad_norm": 1.3447740077972412,
|
| 449 |
+
"learning_rate": 9.955275726883517e-06,
|
| 450 |
+
"loss": 0.3301,
|
| 451 |
+
"step": 63
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 0.2891022021456804,
|
| 455 |
+
"grad_norm": 1.298154592514038,
|
| 456 |
+
"learning_rate": 9.949868360798893e-06,
|
| 457 |
+
"loss": 0.3123,
|
| 458 |
+
"step": 64
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.29361942405420666,
|
| 462 |
+
"grad_norm": 1.3249322175979614,
|
| 463 |
+
"learning_rate": 9.944154131125643e-06,
|
| 464 |
+
"loss": 0.3123,
|
| 465 |
+
"step": 65
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 0.2981366459627329,
|
| 469 |
+
"grad_norm": 1.2478148937225342,
|
| 470 |
+
"learning_rate": 9.938133392113399e-06,
|
| 471 |
+
"loss": 0.3197,
|
| 472 |
+
"step": 66
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 0.3026538678712592,
|
| 476 |
+
"grad_norm": 1.698312520980835,
|
| 477 |
+
"learning_rate": 9.931806517013612e-06,
|
| 478 |
+
"loss": 0.3093,
|
| 479 |
+
"step": 67
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 0.3071710897797854,
|
| 483 |
+
"grad_norm": 1.2709487676620483,
|
| 484 |
+
"learning_rate": 9.925173898056436e-06,
|
| 485 |
+
"loss": 0.3173,
|
| 486 |
+
"step": 68
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 0.3116883116883117,
|
| 490 |
+
"grad_norm": 1.1195577383041382,
|
| 491 |
+
"learning_rate": 9.918235946426389e-06,
|
| 492 |
+
"loss": 0.3161,
|
| 493 |
+
"step": 69
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 0.31620553359683795,
|
| 497 |
+
"grad_norm": 1.6576471328735352,
|
| 498 |
+
"learning_rate": 9.910993092236878e-06,
|
| 499 |
+
"loss": 0.3114,
|
| 500 |
+
"step": 70
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 0.3207227555053642,
|
| 504 |
+
"grad_norm": 1.4347110986709595,
|
| 505 |
+
"learning_rate": 9.903445784503525e-06,
|
| 506 |
+
"loss": 0.3214,
|
| 507 |
+
"step": 71
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 0.32523997741389044,
|
| 511 |
+
"grad_norm": 1.3008146286010742,
|
| 512 |
+
"learning_rate": 9.895594491116336e-06,
|
| 513 |
+
"loss": 0.3054,
|
| 514 |
+
"step": 72
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 0.3297571993224167,
|
| 518 |
+
"grad_norm": 1.2644717693328857,
|
| 519 |
+
"learning_rate": 9.887439698810694e-06,
|
| 520 |
+
"loss": 0.3003,
|
| 521 |
+
"step": 73
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 0.334274421230943,
|
| 525 |
+
"grad_norm": 1.0488579273223877,
|
| 526 |
+
"learning_rate": 9.878981913137178e-06,
|
| 527 |
+
"loss": 0.3126,
|
| 528 |
+
"step": 74
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 0.33879164313946925,
|
| 532 |
+
"grad_norm": 1.3118265867233276,
|
| 533 |
+
"learning_rate": 9.870221658430233e-06,
|
| 534 |
+
"loss": 0.3022,
|
| 535 |
+
"step": 75
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 0.34330886504799546,
|
| 539 |
+
"grad_norm": 1.0711407661437988,
|
| 540 |
+
"learning_rate": 9.861159477775653e-06,
|
| 541 |
+
"loss": 0.3035,
|
| 542 |
+
"step": 76
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 0.34782608695652173,
|
| 546 |
+
"grad_norm": 1.0184773206710815,
|
| 547 |
+
"learning_rate": 9.851795932976919e-06,
|
| 548 |
+
"loss": 0.2932,
|
| 549 |
+
"step": 77
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 0.352343308865048,
|
| 553 |
+
"grad_norm": 1.169956922531128,
|
| 554 |
+
"learning_rate": 9.84213160452037e-06,
|
| 555 |
+
"loss": 0.2986,
|
| 556 |
+
"step": 78
|
| 557 |
+
},
|
| 558 |
+
{
|
| 559 |
+
"epoch": 0.35686053077357427,
|
| 560 |
+
"grad_norm": 1.397139310836792,
|
| 561 |
+
"learning_rate": 9.832167091539215e-06,
|
| 562 |
+
"loss": 0.2981,
|
| 563 |
+
"step": 79
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 0.3613777526821005,
|
| 567 |
+
"grad_norm": 1.212544560432434,
|
| 568 |
+
"learning_rate": 9.821903011776385e-06,
|
| 569 |
+
"loss": 0.3027,
|
| 570 |
+
"step": 80
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 0.36589497459062675,
|
| 574 |
+
"grad_norm": 1.1516971588134766,
|
| 575 |
+
"learning_rate": 9.811340001546252e-06,
|
| 576 |
+
"loss": 0.308,
|
| 577 |
+
"step": 81
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 0.370412196499153,
|
| 581 |
+
"grad_norm": 1.3315784931182861,
|
| 582 |
+
"learning_rate": 9.800478715695165e-06,
|
| 583 |
+
"loss": 0.3042,
|
| 584 |
+
"step": 82
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 0.3749294184076793,
|
| 588 |
+
"grad_norm": 1.2633906602859497,
|
| 589 |
+
"learning_rate": 9.789319827560854e-06,
|
| 590 |
+
"loss": 0.3074,
|
| 591 |
+
"step": 83
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 0.3794466403162055,
|
| 595 |
+
"grad_norm": 1.1565302610397339,
|
| 596 |
+
"learning_rate": 9.777864028930705e-06,
|
| 597 |
+
"loss": 0.2981,
|
| 598 |
+
"step": 84
|
| 599 |
+
},
|
| 600 |
+
{
|
| 601 |
+
"epoch": 0.3839638622247318,
|
| 602 |
+
"grad_norm": 1.307767391204834,
|
| 603 |
+
"learning_rate": 9.766112029998847e-06,
|
| 604 |
+
"loss": 0.289,
|
| 605 |
+
"step": 85
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 0.38848108413325805,
|
| 609 |
+
"grad_norm": 1.3725006580352783,
|
| 610 |
+
"learning_rate": 9.754064559322147e-06,
|
| 611 |
+
"loss": 0.2962,
|
| 612 |
+
"step": 86
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 0.3929983060417843,
|
| 616 |
+
"grad_norm": 1.6594713926315308,
|
| 617 |
+
"learning_rate": 9.741722363775029e-06,
|
| 618 |
+
"loss": 0.2881,
|
| 619 |
+
"step": 87
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"epoch": 0.39751552795031053,
|
| 623 |
+
"grad_norm": 1.1727946996688843,
|
| 624 |
+
"learning_rate": 9.729086208503174e-06,
|
| 625 |
+
"loss": 0.2868,
|
| 626 |
+
"step": 88
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 0.4020327498588368,
|
| 630 |
+
"grad_norm": 1.8756563663482666,
|
| 631 |
+
"learning_rate": 9.716156876876096e-06,
|
| 632 |
+
"loss": 0.2846,
|
| 633 |
+
"step": 89
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 0.40654997176736307,
|
| 637 |
+
"grad_norm": 1.64806067943573,
|
| 638 |
+
"learning_rate": 9.70293517043857e-06,
|
| 639 |
+
"loss": 0.2861,
|
| 640 |
+
"step": 90
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 0.41106719367588934,
|
| 644 |
+
"grad_norm": 1.3863171339035034,
|
| 645 |
+
"learning_rate": 9.689421908860928e-06,
|
| 646 |
+
"loss": 0.3019,
|
| 647 |
+
"step": 91
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 0.4155844155844156,
|
| 651 |
+
"grad_norm": 1.2576581239700317,
|
| 652 |
+
"learning_rate": 9.675617929888271e-06,
|
| 653 |
+
"loss": 0.2908,
|
| 654 |
+
"step": 92
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 0.4201016374929418,
|
| 658 |
+
"grad_norm": 1.0310522317886353,
|
| 659 |
+
"learning_rate": 9.66152408928851e-06,
|
| 660 |
+
"loss": 0.2948,
|
| 661 |
+
"step": 93
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 0.4246188594014681,
|
| 665 |
+
"grad_norm": 1.445862889289856,
|
| 666 |
+
"learning_rate": 9.64714126079933e-06,
|
| 667 |
+
"loss": 0.2793,
|
| 668 |
+
"step": 94
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 0.42913608130999437,
|
| 672 |
+
"grad_norm": 1.3119844198226929,
|
| 673 |
+
"learning_rate": 9.632470336074009e-06,
|
| 674 |
+
"loss": 0.2994,
|
| 675 |
+
"step": 95
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 0.43365330321852064,
|
| 679 |
+
"grad_norm": 1.1884084939956665,
|
| 680 |
+
"learning_rate": 9.617512224626153e-06,
|
| 681 |
+
"loss": 0.3019,
|
| 682 |
+
"step": 96
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 0.43817052512704685,
|
| 686 |
+
"grad_norm": 1.165927529335022,
|
| 687 |
+
"learning_rate": 9.602267853773301e-06,
|
| 688 |
+
"loss": 0.2896,
|
| 689 |
+
"step": 97
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 0.4426877470355731,
|
| 693 |
+
"grad_norm": 1.1704256534576416,
|
| 694 |
+
"learning_rate": 9.586738168579446e-06,
|
| 695 |
+
"loss": 0.2839,
|
| 696 |
+
"step": 98
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 0.4472049689440994,
|
| 700 |
+
"grad_norm": 1.2205328941345215,
|
| 701 |
+
"learning_rate": 9.570924131796437e-06,
|
| 702 |
+
"loss": 0.2755,
|
| 703 |
+
"step": 99
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 0.45172219085262566,
|
| 707 |
+
"grad_norm": 1.3295133113861084,
|
| 708 |
+
"learning_rate": 9.554826723804304e-06,
|
| 709 |
+
"loss": 0.2762,
|
| 710 |
+
"step": 100
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 0.4562394127611519,
|
| 714 |
+
"grad_norm": 1.1630125045776367,
|
| 715 |
+
"learning_rate": 9.538446942550468e-06,
|
| 716 |
+
"loss": 0.276,
|
| 717 |
+
"step": 101
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 0.46075663466967814,
|
| 721 |
+
"grad_norm": 1.326208233833313,
|
| 722 |
+
"learning_rate": 9.521785803487888e-06,
|
| 723 |
+
"loss": 0.2895,
|
| 724 |
+
"step": 102
|
| 725 |
+
},
|
| 726 |
+
{
|
| 727 |
+
"epoch": 0.4652738565782044,
|
| 728 |
+
"grad_norm": 1.2297831773757935,
|
| 729 |
+
"learning_rate": 9.504844339512096e-06,
|
| 730 |
+
"loss": 0.2816,
|
| 731 |
+
"step": 103
|
| 732 |
+
},
|
| 733 |
+
{
|
| 734 |
+
"epoch": 0.4697910784867307,
|
| 735 |
+
"grad_norm": 1.198627233505249,
|
| 736 |
+
"learning_rate": 9.487623600897172e-06,
|
| 737 |
+
"loss": 0.2895,
|
| 738 |
+
"step": 104
|
| 739 |
+
},
|
| 740 |
+
{
|
| 741 |
+
"epoch": 0.4743083003952569,
|
| 742 |
+
"grad_norm": 1.3291443586349487,
|
| 743 |
+
"learning_rate": 9.470124655230627e-06,
|
| 744 |
+
"loss": 0.2816,
|
| 745 |
+
"step": 105
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 0.47882552230378317,
|
| 749 |
+
"grad_norm": 1.138986349105835,
|
| 750 |
+
"learning_rate": 9.452348587347224e-06,
|
| 751 |
+
"loss": 0.2798,
|
| 752 |
+
"step": 106
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 0.48334274421230944,
|
| 756 |
+
"grad_norm": 1.2908552885055542,
|
| 757 |
+
"learning_rate": 9.434296499261719e-06,
|
| 758 |
+
"loss": 0.2819,
|
| 759 |
+
"step": 107
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 0.4878599661208357,
|
| 763 |
+
"grad_norm": 1.4059367179870605,
|
| 764 |
+
"learning_rate": 9.415969510100549e-06,
|
| 765 |
+
"loss": 0.2632,
|
| 766 |
+
"step": 108
|
| 767 |
+
},
|
| 768 |
+
{
|
| 769 |
+
"epoch": 0.4923771880293619,
|
| 770 |
+
"grad_norm": 1.1226232051849365,
|
| 771 |
+
"learning_rate": 9.397368756032445e-06,
|
| 772 |
+
"loss": 0.2864,
|
| 773 |
+
"step": 109
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"epoch": 0.4968944099378882,
|
| 777 |
+
"grad_norm": 1.4748599529266357,
|
| 778 |
+
"learning_rate": 9.378495390198005e-06,
|
| 779 |
+
"loss": 0.2959,
|
| 780 |
+
"step": 110
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"epoch": 0.5014116318464145,
|
| 784 |
+
"grad_norm": 1.2786107063293457,
|
| 785 |
+
"learning_rate": 9.359350582638193e-06,
|
| 786 |
+
"loss": 0.2782,
|
| 787 |
+
"step": 111
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"epoch": 0.5059288537549407,
|
| 791 |
+
"grad_norm": 1.2101049423217773,
|
| 792 |
+
"learning_rate": 9.339935520221816e-06,
|
| 793 |
+
"loss": 0.2729,
|
| 794 |
+
"step": 112
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 0.510446075663467,
|
| 798 |
+
"grad_norm": 1.1712757349014282,
|
| 799 |
+
"learning_rate": 9.32025140657194e-06,
|
| 800 |
+
"loss": 0.2773,
|
| 801 |
+
"step": 113
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 0.5149632975719932,
|
| 805 |
+
"grad_norm": 1.0870469808578491,
|
| 806 |
+
"learning_rate": 9.30029946199127e-06,
|
| 807 |
+
"loss": 0.2823,
|
| 808 |
+
"step": 114
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"epoch": 0.5194805194805194,
|
| 812 |
+
"grad_norm": 1.253859519958496,
|
| 813 |
+
"learning_rate": 9.280080923386501e-06,
|
| 814 |
+
"loss": 0.2708,
|
| 815 |
+
"step": 115
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"epoch": 0.5239977413890458,
|
| 819 |
+
"grad_norm": 1.431169033050537,
|
| 820 |
+
"learning_rate": 9.259597044191635e-06,
|
| 821 |
+
"loss": 0.2769,
|
| 822 |
+
"step": 116
|
| 823 |
+
},
|
| 824 |
+
{
|
| 825 |
+
"epoch": 0.528514963297572,
|
| 826 |
+
"grad_norm": 1.2190595865249634,
|
| 827 |
+
"learning_rate": 9.238849094290279e-06,
|
| 828 |
+
"loss": 0.2746,
|
| 829 |
+
"step": 117
|
| 830 |
+
},
|
| 831 |
+
{
|
| 832 |
+
"epoch": 0.5330321852060983,
|
| 833 |
+
"grad_norm": 1.2797572612762451,
|
| 834 |
+
"learning_rate": 9.217838359936914e-06,
|
| 835 |
+
"loss": 0.2802,
|
| 836 |
+
"step": 118
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 0.5375494071146245,
|
| 840 |
+
"grad_norm": 1.3241180181503296,
|
| 841 |
+
"learning_rate": 9.196566143677157e-06,
|
| 842 |
+
"loss": 0.2781,
|
| 843 |
+
"step": 119
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 0.5420666290231507,
|
| 847 |
+
"grad_norm": 1.4847604036331177,
|
| 848 |
+
"learning_rate": 9.175033764267013e-06,
|
| 849 |
+
"loss": 0.2763,
|
| 850 |
+
"step": 120
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"epoch": 0.546583850931677,
|
| 854 |
+
"grad_norm": 1.0911434888839722,
|
| 855 |
+
"learning_rate": 9.153242556591115e-06,
|
| 856 |
+
"loss": 0.2695,
|
| 857 |
+
"step": 121
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 0.5511010728402033,
|
| 861 |
+
"grad_norm": 1.728698492050171,
|
| 862 |
+
"learning_rate": 9.131193871579975e-06,
|
| 863 |
+
"loss": 0.295,
|
| 864 |
+
"step": 122
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 0.5556182947487295,
|
| 868 |
+
"grad_norm": 1.3457849025726318,
|
| 869 |
+
"learning_rate": 9.108889076126226e-06,
|
| 870 |
+
"loss": 0.2791,
|
| 871 |
+
"step": 123
|
| 872 |
+
},
|
| 873 |
+
{
|
| 874 |
+
"epoch": 0.5601355166572558,
|
| 875 |
+
"grad_norm": 1.5501041412353516,
|
| 876 |
+
"learning_rate": 9.08632955299989e-06,
|
| 877 |
+
"loss": 0.2802,
|
| 878 |
+
"step": 124
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 0.564652738565782,
|
| 882 |
+
"grad_norm": 1.4413127899169922,
|
| 883 |
+
"learning_rate": 9.06351670076265e-06,
|
| 884 |
+
"loss": 0.2672,
|
| 885 |
+
"step": 125
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 0.5691699604743083,
|
| 889 |
+
"grad_norm": 1.0829665660858154,
|
| 890 |
+
"learning_rate": 9.040451933681148e-06,
|
| 891 |
+
"loss": 0.2715,
|
| 892 |
+
"step": 126
|
| 893 |
+
},
|
| 894 |
+
{
|
| 895 |
+
"epoch": 0.5736871823828346,
|
| 896 |
+
"grad_norm": 1.7667239904403687,
|
| 897 |
+
"learning_rate": 9.017136681639307e-06,
|
| 898 |
+
"loss": 0.2797,
|
| 899 |
+
"step": 127
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 0.5782044042913608,
|
| 903 |
+
"grad_norm": 1.6128960847854614,
|
| 904 |
+
"learning_rate": 8.99357239004969e-06,
|
| 905 |
+
"loss": 0.2661,
|
| 906 |
+
"step": 128
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"epoch": 0.5827216261998871,
|
| 910 |
+
"grad_norm": 1.3572909832000732,
|
| 911 |
+
"learning_rate": 8.969760519763891e-06,
|
| 912 |
+
"loss": 0.263,
|
| 913 |
+
"step": 129
|
| 914 |
+
},
|
| 915 |
+
{
|
| 916 |
+
"epoch": 0.5872388481084133,
|
| 917 |
+
"grad_norm": 1.4395819902420044,
|
| 918 |
+
"learning_rate": 8.94570254698197e-06,
|
| 919 |
+
"loss": 0.2768,
|
| 920 |
+
"step": 130
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 0.5917560700169395,
|
| 924 |
+
"grad_norm": 1.1178951263427734,
|
| 925 |
+
"learning_rate": 8.921399963160934e-06,
|
| 926 |
+
"loss": 0.2657,
|
| 927 |
+
"step": 131
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 0.5962732919254659,
|
| 931 |
+
"grad_norm": 1.2729337215423584,
|
| 932 |
+
"learning_rate": 8.89685427492228e-06,
|
| 933 |
+
"loss": 0.2753,
|
| 934 |
+
"step": 132
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 0.6007905138339921,
|
| 938 |
+
"grad_norm": 1.2812000513076782,
|
| 939 |
+
"learning_rate": 8.872067003958597e-06,
|
| 940 |
+
"loss": 0.2859,
|
| 941 |
+
"step": 133
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 0.6053077357425184,
|
| 945 |
+
"grad_norm": 1.1353107690811157,
|
| 946 |
+
"learning_rate": 8.847039686939218e-06,
|
| 947 |
+
"loss": 0.2686,
|
| 948 |
+
"step": 134
|
| 949 |
+
},
|
| 950 |
+
{
|
| 951 |
+
"epoch": 0.6098249576510446,
|
| 952 |
+
"grad_norm": 1.2572075128555298,
|
| 953 |
+
"learning_rate": 8.82177387541496e-06,
|
| 954 |
+
"loss": 0.2625,
|
| 955 |
+
"step": 135
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 0.6143421795595708,
|
| 959 |
+
"grad_norm": 1.0186498165130615,
|
| 960 |
+
"learning_rate": 8.796271135721944e-06,
|
| 961 |
+
"loss": 0.2734,
|
| 962 |
+
"step": 136
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 0.6188594014680971,
|
| 966 |
+
"grad_norm": 1.2548401355743408,
|
| 967 |
+
"learning_rate": 8.770533048884483e-06,
|
| 968 |
+
"loss": 0.2627,
|
| 969 |
+
"step": 137
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 0.6233766233766234,
|
| 973 |
+
"grad_norm": 1.1643537282943726,
|
| 974 |
+
"learning_rate": 8.744561210517067e-06,
|
| 975 |
+
"loss": 0.2738,
|
| 976 |
+
"step": 138
|
| 977 |
+
},
|
| 978 |
+
{
|
| 979 |
+
"epoch": 0.6278938452851497,
|
| 980 |
+
"grad_norm": 1.2717437744140625,
|
| 981 |
+
"learning_rate": 8.71835723072545e-06,
|
| 982 |
+
"loss": 0.2652,
|
| 983 |
+
"step": 139
|
| 984 |
+
},
|
| 985 |
+
{
|
| 986 |
+
"epoch": 0.6324110671936759,
|
| 987 |
+
"grad_norm": 1.2601284980773926,
|
| 988 |
+
"learning_rate": 8.691922734006828e-06,
|
| 989 |
+
"loss": 0.2764,
|
| 990 |
+
"step": 140
|
| 991 |
+
},
|
| 992 |
+
{
|
| 993 |
+
"epoch": 0.6369282891022021,
|
| 994 |
+
"grad_norm": 1.122058391571045,
|
| 995 |
+
"learning_rate": 8.665259359149132e-06,
|
| 996 |
+
"loss": 0.2587,
|
| 997 |
+
"step": 141
|
| 998 |
+
},
|
| 999 |
+
{
|
| 1000 |
+
"epoch": 0.6414455110107284,
|
| 1001 |
+
"grad_norm": 1.0772087574005127,
|
| 1002 |
+
"learning_rate": 8.638368759129433e-06,
|
| 1003 |
+
"loss": 0.2635,
|
| 1004 |
+
"step": 142
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 0.6459627329192547,
|
| 1008 |
+
"grad_norm": 1.126037836074829,
|
| 1009 |
+
"learning_rate": 8.611252601011457e-06,
|
| 1010 |
+
"loss": 0.2555,
|
| 1011 |
+
"step": 143
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 0.6504799548277809,
|
| 1015 |
+
"grad_norm": 1.1625555753707886,
|
| 1016 |
+
"learning_rate": 8.583912565842258e-06,
|
| 1017 |
+
"loss": 0.2843,
|
| 1018 |
+
"step": 144
|
| 1019 |
+
},
|
| 1020 |
+
{
|
| 1021 |
+
"epoch": 0.6549971767363072,
|
| 1022 |
+
"grad_norm": 1.1770436763763428,
|
| 1023 |
+
"learning_rate": 8.556350348547978e-06,
|
| 1024 |
+
"loss": 0.2762,
|
| 1025 |
+
"step": 145
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 0.6595143986448334,
|
| 1029 |
+
"grad_norm": 1.0764650106430054,
|
| 1030 |
+
"learning_rate": 8.528567657828785e-06,
|
| 1031 |
+
"loss": 0.2637,
|
| 1032 |
+
"step": 146
|
| 1033 |
+
},
|
| 1034 |
+
{
|
| 1035 |
+
"epoch": 0.6640316205533597,
|
| 1036 |
+
"grad_norm": 1.4244846105575562,
|
| 1037 |
+
"learning_rate": 8.500566216052948e-06,
|
| 1038 |
+
"loss": 0.2737,
|
| 1039 |
+
"step": 147
|
| 1040 |
+
},
|
| 1041 |
+
{
|
| 1042 |
+
"epoch": 0.668548842461886,
|
| 1043 |
+
"grad_norm": 1.4140225648880005,
|
| 1044 |
+
"learning_rate": 8.472347759150044e-06,
|
| 1045 |
+
"loss": 0.2623,
|
| 1046 |
+
"step": 148
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 0.6730660643704122,
|
| 1050 |
+
"grad_norm": 1.120361328125,
|
| 1051 |
+
"learning_rate": 8.443914036503356e-06,
|
| 1052 |
+
"loss": 0.2708,
|
| 1053 |
+
"step": 149
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 0.6775832862789385,
|
| 1057 |
+
"grad_norm": 1.135714054107666,
|
| 1058 |
+
"learning_rate": 8.415266810841412e-06,
|
| 1059 |
+
"loss": 0.258,
|
| 1060 |
+
"step": 150
|
| 1061 |
+
},
|
| 1062 |
+
{
|
| 1063 |
+
"epoch": 0.6821005081874647,
|
| 1064 |
+
"grad_norm": 1.1812201738357544,
|
| 1065 |
+
"learning_rate": 8.386407858128707e-06,
|
| 1066 |
+
"loss": 0.2674,
|
| 1067 |
+
"step": 151
|
| 1068 |
+
},
|
| 1069 |
+
{
|
| 1070 |
+
"epoch": 0.6866177300959909,
|
| 1071 |
+
"grad_norm": 1.1621167659759521,
|
| 1072 |
+
"learning_rate": 8.357338967455605e-06,
|
| 1073 |
+
"loss": 0.2702,
|
| 1074 |
+
"step": 152
|
| 1075 |
+
},
|
| 1076 |
+
{
|
| 1077 |
+
"epoch": 0.6911349520045172,
|
| 1078 |
+
"grad_norm": 1.1746842861175537,
|
| 1079 |
+
"learning_rate": 8.32806194092743e-06,
|
| 1080 |
+
"loss": 0.2619,
|
| 1081 |
+
"step": 153
|
| 1082 |
+
},
|
| 1083 |
+
{
|
| 1084 |
+
"epoch": 0.6956521739130435,
|
| 1085 |
+
"grad_norm": 1.1322088241577148,
|
| 1086 |
+
"learning_rate": 8.298578593552737e-06,
|
| 1087 |
+
"loss": 0.2534,
|
| 1088 |
+
"step": 154
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 0.7001693958215698,
|
| 1092 |
+
"grad_norm": 1.170893669128418,
|
| 1093 |
+
"learning_rate": 8.268890753130794e-06,
|
| 1094 |
+
"loss": 0.2698,
|
| 1095 |
+
"step": 155
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 0.704686617730096,
|
| 1099 |
+
"grad_norm": 1.2007019519805908,
|
| 1100 |
+
"learning_rate": 8.239000260138277e-06,
|
| 1101 |
+
"loss": 0.2599,
|
| 1102 |
+
"step": 156
|
| 1103 |
+
},
|
| 1104 |
+
{
|
| 1105 |
+
"epoch": 0.7092038396386222,
|
| 1106 |
+
"grad_norm": 1.399940013885498,
|
| 1107 |
+
"learning_rate": 8.208908967615159e-06,
|
| 1108 |
+
"loss": 0.2678,
|
| 1109 |
+
"step": 157
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 0.7137210615471485,
|
| 1113 |
+
"grad_norm": 1.1865079402923584,
|
| 1114 |
+
"learning_rate": 8.178618741049841e-06,
|
| 1115 |
+
"loss": 0.2712,
|
| 1116 |
+
"step": 158
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 0.7182382834556748,
|
| 1120 |
+
"grad_norm": 1.3443241119384766,
|
| 1121 |
+
"learning_rate": 8.148131458263499e-06,
|
| 1122 |
+
"loss": 0.2635,
|
| 1123 |
+
"step": 159
|
| 1124 |
+
},
|
| 1125 |
+
{
|
| 1126 |
+
"epoch": 0.722755505364201,
|
| 1127 |
+
"grad_norm": 1.0922118425369263,
|
| 1128 |
+
"learning_rate": 8.117449009293668e-06,
|
| 1129 |
+
"loss": 0.2688,
|
| 1130 |
+
"step": 160
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 0.7272727272727273,
|
| 1134 |
+
"grad_norm": 1.2795259952545166,
|
| 1135 |
+
"learning_rate": 8.086573296277078e-06,
|
| 1136 |
+
"loss": 0.2723,
|
| 1137 |
+
"step": 161
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 0.7317899491812535,
|
| 1141 |
+
"grad_norm": 1.0830557346343994,
|
| 1142 |
+
"learning_rate": 8.055506233331718e-06,
|
| 1143 |
+
"loss": 0.262,
|
| 1144 |
+
"step": 162
|
| 1145 |
+
},
|
| 1146 |
+
{
|
| 1147 |
+
"epoch": 0.7363071710897798,
|
| 1148 |
+
"grad_norm": 1.2854331731796265,
|
| 1149 |
+
"learning_rate": 8.024249746438189e-06,
|
| 1150 |
+
"loss": 0.2654,
|
| 1151 |
+
"step": 163
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 0.740824392998306,
|
| 1155 |
+
"grad_norm": 1.1591980457305908,
|
| 1156 |
+
"learning_rate": 7.99280577332029e-06,
|
| 1157 |
+
"loss": 0.2656,
|
| 1158 |
+
"step": 164
|
| 1159 |
+
},
|
| 1160 |
+
{
|
| 1161 |
+
"epoch": 0.7453416149068323,
|
| 1162 |
+
"grad_norm": 1.0863441228866577,
|
| 1163 |
+
"learning_rate": 7.961176263324902e-06,
|
| 1164 |
+
"loss": 0.2662,
|
| 1165 |
+
"step": 165
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 0.7498588368153586,
|
| 1169 |
+
"grad_norm": 1.2065740823745728,
|
| 1170 |
+
"learning_rate": 7.929363177301124e-06,
|
| 1171 |
+
"loss": 0.257,
|
| 1172 |
+
"step": 166
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 0.7543760587238848,
|
| 1176 |
+
"grad_norm": 1.1085830926895142,
|
| 1177 |
+
"learning_rate": 7.897368487478733e-06,
|
| 1178 |
+
"loss": 0.2547,
|
| 1179 |
+
"step": 167
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 0.758893280632411,
|
| 1183 |
+
"grad_norm": 1.1087521314620972,
|
| 1184 |
+
"learning_rate": 7.865194177345894e-06,
|
| 1185 |
+
"loss": 0.2605,
|
| 1186 |
+
"step": 168
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 0.7634105025409373,
|
| 1190 |
+
"grad_norm": 1.2097454071044922,
|
| 1191 |
+
"learning_rate": 7.832842241526212e-06,
|
| 1192 |
+
"loss": 0.2513,
|
| 1193 |
+
"step": 169
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 0.7679277244494636,
|
| 1197 |
+
"grad_norm": 1.1733957529067993,
|
| 1198 |
+
"learning_rate": 7.800314685655072e-06,
|
| 1199 |
+
"loss": 0.27,
|
| 1200 |
+
"step": 170
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 0.7724449463579899,
|
| 1204 |
+
"grad_norm": 1.1870712041854858,
|
| 1205 |
+
"learning_rate": 7.767613526255296e-06,
|
| 1206 |
+
"loss": 0.257,
|
| 1207 |
+
"step": 171
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 0.7769621682665161,
|
| 1211 |
+
"grad_norm": 1.3030035495758057,
|
| 1212 |
+
"learning_rate": 7.734740790612137e-06,
|
| 1213 |
+
"loss": 0.2711,
|
| 1214 |
+
"step": 172
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 0.7814793901750423,
|
| 1218 |
+
"grad_norm": 1.1683826446533203,
|
| 1219 |
+
"learning_rate": 7.70169851664759e-06,
|
| 1220 |
+
"loss": 0.2395,
|
| 1221 |
+
"step": 173
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 0.7859966120835686,
|
| 1225 |
+
"grad_norm": 1.303270936012268,
|
| 1226 |
+
"learning_rate": 7.668488752794067e-06,
|
| 1227 |
+
"loss": 0.2784,
|
| 1228 |
+
"step": 174
|
| 1229 |
+
},
|
| 1230 |
+
{
|
| 1231 |
+
"epoch": 0.7905138339920948,
|
| 1232 |
+
"grad_norm": 1.0825122594833374,
|
| 1233 |
+
"learning_rate": 7.635113557867395e-06,
|
| 1234 |
+
"loss": 0.2601,
|
| 1235 |
+
"step": 175
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 0.7950310559006211,
|
| 1239 |
+
"grad_norm": 1.1751047372817993,
|
| 1240 |
+
"learning_rate": 7.6015750009391776e-06,
|
| 1241 |
+
"loss": 0.2502,
|
| 1242 |
+
"step": 176
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 0.7995482778091474,
|
| 1246 |
+
"grad_norm": 1.2013722658157349,
|
| 1247 |
+
"learning_rate": 7.5678751612085344e-06,
|
| 1248 |
+
"loss": 0.2479,
|
| 1249 |
+
"step": 177
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 0.8040654997176736,
|
| 1253 |
+
"grad_norm": 1.1240850687026978,
|
| 1254 |
+
"learning_rate": 7.5340161278732e-06,
|
| 1255 |
+
"loss": 0.2606,
|
| 1256 |
+
"step": 178
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 0.8085827216261999,
|
| 1260 |
+
"grad_norm": 1.260583758354187,
|
| 1261 |
+
"learning_rate": 7.500000000000001e-06,
|
| 1262 |
+
"loss": 0.2615,
|
| 1263 |
+
"step": 179
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 0.8130999435347261,
|
| 1267 |
+
"grad_norm": 1.1350092887878418,
|
| 1268 |
+
"learning_rate": 7.465828886394729e-06,
|
| 1269 |
+
"loss": 0.2564,
|
| 1270 |
+
"step": 180
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 0.8176171654432524,
|
| 1274 |
+
"grad_norm": 1.0673713684082031,
|
| 1275 |
+
"learning_rate": 7.431504905471407e-06,
|
| 1276 |
+
"loss": 0.2576,
|
| 1277 |
+
"step": 181
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 0.8221343873517787,
|
| 1281 |
+
"grad_norm": 1.3356529474258423,
|
| 1282 |
+
"learning_rate": 7.397030185120962e-06,
|
| 1283 |
+
"loss": 0.2515,
|
| 1284 |
+
"step": 182
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 0.8266516092603049,
|
| 1288 |
+
"grad_norm": 1.1309804916381836,
|
| 1289 |
+
"learning_rate": 7.362406862579299e-06,
|
| 1290 |
+
"loss": 0.255,
|
| 1291 |
+
"step": 183
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 0.8311688311688312,
|
| 1295 |
+
"grad_norm": 1.4243286848068237,
|
| 1296 |
+
"learning_rate": 7.327637084294818e-06,
|
| 1297 |
+
"loss": 0.2559,
|
| 1298 |
+
"step": 184
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 0.8356860530773574,
|
| 1302 |
+
"grad_norm": 1.2205266952514648,
|
| 1303 |
+
"learning_rate": 7.29272300579533e-06,
|
| 1304 |
+
"loss": 0.2447,
|
| 1305 |
+
"step": 185
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 0.8402032749858837,
|
| 1309 |
+
"grad_norm": 1.2443662881851196,
|
| 1310 |
+
"learning_rate": 7.257666791554448e-06,
|
| 1311 |
+
"loss": 0.261,
|
| 1312 |
+
"step": 186
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"epoch": 0.84472049689441,
|
| 1316 |
+
"grad_norm": 1.1030696630477905,
|
| 1317 |
+
"learning_rate": 7.22247061485738e-06,
|
| 1318 |
+
"loss": 0.2523,
|
| 1319 |
+
"step": 187
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 0.8492377188029362,
|
| 1323 |
+
"grad_norm": 1.026666522026062,
|
| 1324 |
+
"learning_rate": 7.187136657666208e-06,
|
| 1325 |
+
"loss": 0.2547,
|
| 1326 |
+
"step": 188
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 0.8537549407114624,
|
| 1330 |
+
"grad_norm": 1.1070029735565186,
|
| 1331 |
+
"learning_rate": 7.151667110484626e-06,
|
| 1332 |
+
"loss": 0.2435,
|
| 1333 |
+
"step": 189
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 0.8582721626199887,
|
| 1337 |
+
"grad_norm": 1.2672240734100342,
|
| 1338 |
+
"learning_rate": 7.1160641722221255e-06,
|
| 1339 |
+
"loss": 0.2455,
|
| 1340 |
+
"step": 190
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 0.862789384528515,
|
| 1344 |
+
"grad_norm": 1.2531083822250366,
|
| 1345 |
+
"learning_rate": 7.080330050057687e-06,
|
| 1346 |
+
"loss": 0.2641,
|
| 1347 |
+
"step": 191
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 0.8673066064370413,
|
| 1351 |
+
"grad_norm": 1.1208117008209229,
|
| 1352 |
+
"learning_rate": 7.044466959302945e-06,
|
| 1353 |
+
"loss": 0.2403,
|
| 1354 |
+
"step": 192
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 0.8718238283455675,
|
| 1358 |
+
"grad_norm": 1.1949052810668945,
|
| 1359 |
+
"learning_rate": 7.008477123264849e-06,
|
| 1360 |
+
"loss": 0.2623,
|
| 1361 |
+
"step": 193
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 0.8763410502540937,
|
| 1365 |
+
"grad_norm": 1.0727616548538208,
|
| 1366 |
+
"learning_rate": 6.972362773107832e-06,
|
| 1367 |
+
"loss": 0.2471,
|
| 1368 |
+
"step": 194
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 0.88085827216262,
|
| 1372 |
+
"grad_norm": 1.0367906093597412,
|
| 1373 |
+
"learning_rate": 6.936126147715494e-06,
|
| 1374 |
+
"loss": 0.2489,
|
| 1375 |
+
"step": 195
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 0.8853754940711462,
|
| 1379 |
+
"grad_norm": 1.0889986753463745,
|
| 1380 |
+
"learning_rate": 6.8997694935518e-06,
|
| 1381 |
+
"loss": 0.2445,
|
| 1382 |
+
"step": 196
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 0.8898927159796725,
|
| 1386 |
+
"grad_norm": 1.0107128620147705,
|
| 1387 |
+
"learning_rate": 6.863295064521816e-06,
|
| 1388 |
+
"loss": 0.248,
|
| 1389 |
+
"step": 197
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 0.8944099378881988,
|
| 1393 |
+
"grad_norm": 1.0799349546432495,
|
| 1394 |
+
"learning_rate": 6.8267051218319766e-06,
|
| 1395 |
+
"loss": 0.2525,
|
| 1396 |
+
"step": 198
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 0.898927159796725,
|
| 1400 |
+
"grad_norm": 1.077852487564087,
|
| 1401 |
+
"learning_rate": 6.7900019338499005e-06,
|
| 1402 |
+
"loss": 0.2512,
|
| 1403 |
+
"step": 199
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 0.9034443817052513,
|
| 1407 |
+
"grad_norm": 1.0199133157730103,
|
| 1408 |
+
"learning_rate": 6.753187775963773e-06,
|
| 1409 |
+
"loss": 0.252,
|
| 1410 |
+
"step": 200
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 0.9079616036137775,
|
| 1414 |
+
"grad_norm": 1.204663872718811,
|
| 1415 |
+
"learning_rate": 6.716264930441279e-06,
|
| 1416 |
+
"loss": 0.261,
|
| 1417 |
+
"step": 201
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 0.9124788255223037,
|
| 1421 |
+
"grad_norm": 1.2283254861831665,
|
| 1422 |
+
"learning_rate": 6.6792356862881144e-06,
|
| 1423 |
+
"loss": 0.2593,
|
| 1424 |
+
"step": 202
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 0.9169960474308301,
|
| 1428 |
+
"grad_norm": 1.0611567497253418,
|
| 1429 |
+
"learning_rate": 6.6421023391060845e-06,
|
| 1430 |
+
"loss": 0.2514,
|
| 1431 |
+
"step": 203
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 0.9215132693393563,
|
| 1435 |
+
"grad_norm": 1.2056411504745483,
|
| 1436 |
+
"learning_rate": 6.60486719095079e-06,
|
| 1437 |
+
"loss": 0.2649,
|
| 1438 |
+
"step": 204
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 0.9260304912478825,
|
| 1442 |
+
"grad_norm": 1.1966333389282227,
|
| 1443 |
+
"learning_rate": 6.567532550188908e-06,
|
| 1444 |
+
"loss": 0.2611,
|
| 1445 |
+
"step": 205
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 0.9305477131564088,
|
| 1449 |
+
"grad_norm": 1.2905259132385254,
|
| 1450 |
+
"learning_rate": 6.530100731355089e-06,
|
| 1451 |
+
"loss": 0.255,
|
| 1452 |
+
"step": 206
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"epoch": 0.935064935064935,
|
| 1456 |
+
"grad_norm": 1.037772536277771,
|
| 1457 |
+
"learning_rate": 6.492574055008474e-06,
|
| 1458 |
+
"loss": 0.2495,
|
| 1459 |
+
"step": 207
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 0.9395821569734614,
|
| 1463 |
+
"grad_norm": 1.1251964569091797,
|
| 1464 |
+
"learning_rate": 6.454954847588824e-06,
|
| 1465 |
+
"loss": 0.2407,
|
| 1466 |
+
"step": 208
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 0.9440993788819876,
|
| 1470 |
+
"grad_norm": 1.0818254947662354,
|
| 1471 |
+
"learning_rate": 6.417245441272299e-06,
|
| 1472 |
+
"loss": 0.2615,
|
| 1473 |
+
"step": 209
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 0.9486166007905138,
|
| 1477 |
+
"grad_norm": 1.1475183963775635,
|
| 1478 |
+
"learning_rate": 6.3794481738268765e-06,
|
| 1479 |
+
"loss": 0.2493,
|
| 1480 |
+
"step": 210
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 0.9531338226990401,
|
| 1484 |
+
"grad_norm": 1.268548846244812,
|
| 1485 |
+
"learning_rate": 6.341565388467425e-06,
|
| 1486 |
+
"loss": 0.2629,
|
| 1487 |
+
"step": 211
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 0.9576510446075663,
|
| 1491 |
+
"grad_norm": 1.215813159942627,
|
| 1492 |
+
"learning_rate": 6.30359943371043e-06,
|
| 1493 |
+
"loss": 0.2462,
|
| 1494 |
+
"step": 212
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 0.9621682665160926,
|
| 1498 |
+
"grad_norm": 1.0642273426055908,
|
| 1499 |
+
"learning_rate": 6.265552663228411e-06,
|
| 1500 |
+
"loss": 0.2419,
|
| 1501 |
+
"step": 213
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 0.9666854884246189,
|
| 1505 |
+
"grad_norm": 1.0285767316818237,
|
| 1506 |
+
"learning_rate": 6.227427435703997e-06,
|
| 1507 |
+
"loss": 0.2474,
|
| 1508 |
+
"step": 214
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 0.9712027103331451,
|
| 1512 |
+
"grad_norm": 1.1549084186553955,
|
| 1513 |
+
"learning_rate": 6.189226114683708e-06,
|
| 1514 |
+
"loss": 0.2716,
|
| 1515 |
+
"step": 215
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 0.9757199322416714,
|
| 1519 |
+
"grad_norm": 1.1253952980041504,
|
| 1520 |
+
"learning_rate": 6.150951068431424e-06,
|
| 1521 |
+
"loss": 0.2519,
|
| 1522 |
+
"step": 216
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 0.9802371541501976,
|
| 1526 |
+
"grad_norm": 1.1693516969680786,
|
| 1527 |
+
"learning_rate": 6.112604669781572e-06,
|
| 1528 |
+
"loss": 0.2455,
|
| 1529 |
+
"step": 217
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 0.9847543760587238,
|
| 1533 |
+
"grad_norm": 1.1159557104110718,
|
| 1534 |
+
"learning_rate": 6.0741892959920205e-06,
|
| 1535 |
+
"loss": 0.2508,
|
| 1536 |
+
"step": 218
|
| 1537 |
+
},
|
| 1538 |
+
{
|
| 1539 |
+
"epoch": 0.9892715979672502,
|
| 1540 |
+
"grad_norm": 1.0758064985275269,
|
| 1541 |
+
"learning_rate": 6.035707328596698e-06,
|
| 1542 |
+
"loss": 0.2433,
|
| 1543 |
+
"step": 219
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"epoch": 0.9937888198757764,
|
| 1547 |
+
"grad_norm": 1.1275575160980225,
|
| 1548 |
+
"learning_rate": 5.997161153257963e-06,
|
| 1549 |
+
"loss": 0.2481,
|
| 1550 |
+
"step": 220
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 0.9983060417843026,
|
| 1554 |
+
"grad_norm": 1.002582311630249,
|
| 1555 |
+
"learning_rate": 5.958553159618693e-06,
|
| 1556 |
+
"loss": 0.2419,
|
| 1557 |
+
"step": 221
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 1.0,
|
| 1561 |
+
"grad_norm": 1.4576319456100464,
|
| 1562 |
+
"learning_rate": 5.919885741154155e-06,
|
| 1563 |
+
"loss": 0.2324,
|
| 1564 |
+
"step": 222
|
| 1565 |
+
},
|
| 1566 |
+
{
|
| 1567 |
+
"epoch": 1.0045172219085263,
|
| 1568 |
+
"grad_norm": 1.2537062168121338,
|
| 1569 |
+
"learning_rate": 5.88116129502361e-06,
|
| 1570 |
+
"loss": 0.2127,
|
| 1571 |
+
"step": 223
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 1.0090344438170524,
|
| 1575 |
+
"grad_norm": 1.2806092500686646,
|
| 1576 |
+
"learning_rate": 5.842382221921711e-06,
|
| 1577 |
+
"loss": 0.2252,
|
| 1578 |
+
"step": 224
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 1.0135516657255788,
|
| 1582 |
+
"grad_norm": 1.0919123888015747,
|
| 1583 |
+
"learning_rate": 5.803550925929673e-06,
|
| 1584 |
+
"loss": 0.2338,
|
| 1585 |
+
"step": 225
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"epoch": 1.018068887634105,
|
| 1589 |
+
"grad_norm": 1.1968063116073608,
|
| 1590 |
+
"learning_rate": 5.764669814366231e-06,
|
| 1591 |
+
"loss": 0.233,
|
| 1592 |
+
"step": 226
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 1.0225861095426312,
|
| 1596 |
+
"grad_norm": 1.107119083404541,
|
| 1597 |
+
"learning_rate": 5.725741297638399e-06,
|
| 1598 |
+
"loss": 0.2295,
|
| 1599 |
+
"step": 227
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 1.0271033314511575,
|
| 1603 |
+
"grad_norm": 1.159583330154419,
|
| 1604 |
+
"learning_rate": 5.686767789092041e-06,
|
| 1605 |
+
"loss": 0.2257,
|
| 1606 |
+
"step": 228
|
| 1607 |
+
},
|
| 1608 |
+
{
|
| 1609 |
+
"epoch": 1.0316205533596838,
|
| 1610 |
+
"grad_norm": 1.1898163557052612,
|
| 1611 |
+
"learning_rate": 5.647751704862263e-06,
|
| 1612 |
+
"loss": 0.2233,
|
| 1613 |
+
"step": 229
|
| 1614 |
+
},
|
| 1615 |
+
{
|
| 1616 |
+
"epoch": 1.0361377752682102,
|
| 1617 |
+
"grad_norm": 1.1058701276779175,
|
| 1618 |
+
"learning_rate": 5.608695463723614e-06,
|
| 1619 |
+
"loss": 0.2251,
|
| 1620 |
+
"step": 230
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 1.0406549971767363,
|
| 1624 |
+
"grad_norm": 1.0528849363327026,
|
| 1625 |
+
"learning_rate": 5.56960148694014e-06,
|
| 1626 |
+
"loss": 0.2254,
|
| 1627 |
+
"step": 231
|
| 1628 |
+
},
|
| 1629 |
+
{
|
| 1630 |
+
"epoch": 1.0451722190852626,
|
| 1631 |
+
"grad_norm": 1.2264484167099,
|
| 1632 |
+
"learning_rate": 5.530472198115291e-06,
|
| 1633 |
+
"loss": 0.2197,
|
| 1634 |
+
"step": 232
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 1.049689440993789,
|
| 1638 |
+
"grad_norm": 1.340097427368164,
|
| 1639 |
+
"learning_rate": 5.4913100230416536e-06,
|
| 1640 |
+
"loss": 0.2176,
|
| 1641 |
+
"step": 233
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 1.054206662902315,
|
| 1645 |
+
"grad_norm": 1.254300594329834,
|
| 1646 |
+
"learning_rate": 5.45211738955058e-06,
|
| 1647 |
+
"loss": 0.2235,
|
| 1648 |
+
"step": 234
|
| 1649 |
+
},
|
| 1650 |
+
{
|
| 1651 |
+
"epoch": 1.0587238848108413,
|
| 1652 |
+
"grad_norm": 1.0747822523117065,
|
| 1653 |
+
"learning_rate": 5.412896727361663e-06,
|
| 1654 |
+
"loss": 0.2239,
|
| 1655 |
+
"step": 235
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 1.0632411067193677,
|
| 1659 |
+
"grad_norm": 1.1188759803771973,
|
| 1660 |
+
"learning_rate": 5.373650467932122e-06,
|
| 1661 |
+
"loss": 0.2265,
|
| 1662 |
+
"step": 236
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 1.0677583286278938,
|
| 1666 |
+
"grad_norm": 1.2266041040420532,
|
| 1667 |
+
"learning_rate": 5.334381044306057e-06,
|
| 1668 |
+
"loss": 0.2227,
|
| 1669 |
+
"step": 237
|
| 1670 |
+
},
|
| 1671 |
+
{
|
| 1672 |
+
"epoch": 1.07227555053642,
|
| 1673 |
+
"grad_norm": 1.2001854181289673,
|
| 1674 |
+
"learning_rate": 5.2950908909636144e-06,
|
| 1675 |
+
"loss": 0.2341,
|
| 1676 |
+
"step": 238
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 1.0767927724449464,
|
| 1680 |
+
"grad_norm": 1.0783330202102661,
|
| 1681 |
+
"learning_rate": 5.255782443670068e-06,
|
| 1682 |
+
"loss": 0.2164,
|
| 1683 |
+
"step": 239
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 1.0813099943534725,
|
| 1687 |
+
"grad_norm": 1.0817006826400757,
|
| 1688 |
+
"learning_rate": 5.216458139324806e-06,
|
| 1689 |
+
"loss": 0.2276,
|
| 1690 |
+
"step": 240
|
| 1691 |
+
},
|
| 1692 |
+
{
|
| 1693 |
+
"epoch": 1.0858272162619989,
|
| 1694 |
+
"grad_norm": 1.0655776262283325,
|
| 1695 |
+
"learning_rate": 5.177120415810271e-06,
|
| 1696 |
+
"loss": 0.2209,
|
| 1697 |
+
"step": 241
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 1.0903444381705252,
|
| 1701 |
+
"grad_norm": 1.110958456993103,
|
| 1702 |
+
"learning_rate": 5.137771711840811e-06,
|
| 1703 |
+
"loss": 0.2211,
|
| 1704 |
+
"step": 242
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 1.0948616600790513,
|
| 1708 |
+
"grad_norm": 1.1167265176773071,
|
| 1709 |
+
"learning_rate": 5.098414466811504e-06,
|
| 1710 |
+
"loss": 0.2099,
|
| 1711 |
+
"step": 243
|
| 1712 |
+
},
|
| 1713 |
+
{
|
| 1714 |
+
"epoch": 1.0993788819875776,
|
| 1715 |
+
"grad_norm": 1.071454405784607,
|
| 1716 |
+
"learning_rate": 5.059051120646924e-06,
|
| 1717 |
+
"loss": 0.221,
|
| 1718 |
+
"step": 244
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 1.103896103896104,
|
| 1722 |
+
"grad_norm": 1.215519666671753,
|
| 1723 |
+
"learning_rate": 5.019684113649877e-06,
|
| 1724 |
+
"loss": 0.2211,
|
| 1725 |
+
"step": 245
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 1.1084133258046303,
|
| 1729 |
+
"grad_norm": 0.9633514881134033,
|
| 1730 |
+
"learning_rate": 4.980315886350125e-06,
|
| 1731 |
+
"loss": 0.2177,
|
| 1732 |
+
"step": 246
|
| 1733 |
+
},
|
| 1734 |
+
{
|
| 1735 |
+
"epoch": 1.1129305477131564,
|
| 1736 |
+
"grad_norm": 1.0305547714233398,
|
| 1737 |
+
"learning_rate": 4.940948879353078e-06,
|
| 1738 |
+
"loss": 0.2119,
|
| 1739 |
+
"step": 247
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 1.1174477696216827,
|
| 1743 |
+
"grad_norm": 1.1536132097244263,
|
| 1744 |
+
"learning_rate": 4.9015855331884984e-06,
|
| 1745 |
+
"loss": 0.2244,
|
| 1746 |
+
"step": 248
|
| 1747 |
+
},
|
| 1748 |
+
{
|
| 1749 |
+
"epoch": 1.121964991530209,
|
| 1750 |
+
"grad_norm": 1.081312894821167,
|
| 1751 |
+
"learning_rate": 4.862228288159191e-06,
|
| 1752 |
+
"loss": 0.2175,
|
| 1753 |
+
"step": 249
|
| 1754 |
+
},
|
| 1755 |
+
{
|
| 1756 |
+
"epoch": 1.1264822134387351,
|
| 1757 |
+
"grad_norm": 1.0267103910446167,
|
| 1758 |
+
"learning_rate": 4.822879584189732e-06,
|
| 1759 |
+
"loss": 0.2172,
|
| 1760 |
+
"step": 250
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 1.1309994353472614,
|
| 1764 |
+
"grad_norm": 1.30397367477417,
|
| 1765 |
+
"learning_rate": 4.783541860675195e-06,
|
| 1766 |
+
"loss": 0.2163,
|
| 1767 |
+
"step": 251
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 1.1355166572557878,
|
| 1771 |
+
"grad_norm": 1.1941206455230713,
|
| 1772 |
+
"learning_rate": 4.744217556329935e-06,
|
| 1773 |
+
"loss": 0.224,
|
| 1774 |
+
"step": 252
|
| 1775 |
+
},
|
| 1776 |
+
{
|
| 1777 |
+
"epoch": 1.1400338791643139,
|
| 1778 |
+
"grad_norm": 1.2436816692352295,
|
| 1779 |
+
"learning_rate": 4.704909109036387e-06,
|
| 1780 |
+
"loss": 0.2138,
|
| 1781 |
+
"step": 253
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 1.1445511010728402,
|
| 1785 |
+
"grad_norm": 1.1815208196640015,
|
| 1786 |
+
"learning_rate": 4.6656189556939446e-06,
|
| 1787 |
+
"loss": 0.22,
|
| 1788 |
+
"step": 254
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 1.1490683229813665,
|
| 1792 |
+
"grad_norm": 1.3059700727462769,
|
| 1793 |
+
"learning_rate": 4.626349532067879e-06,
|
| 1794 |
+
"loss": 0.2353,
|
| 1795 |
+
"step": 255
|
| 1796 |
+
},
|
| 1797 |
+
{
|
| 1798 |
+
"epoch": 1.1535855448898926,
|
| 1799 |
+
"grad_norm": 1.0804404020309448,
|
| 1800 |
+
"learning_rate": 4.587103272638339e-06,
|
| 1801 |
+
"loss": 0.224,
|
| 1802 |
+
"step": 256
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 1.158102766798419,
|
| 1806 |
+
"grad_norm": 1.0924152135849,
|
| 1807 |
+
"learning_rate": 4.5478826104494225e-06,
|
| 1808 |
+
"loss": 0.2178,
|
| 1809 |
+
"step": 257
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 1.1626199887069453,
|
| 1813 |
+
"grad_norm": 1.2975399494171143,
|
| 1814 |
+
"learning_rate": 4.508689976958348e-06,
|
| 1815 |
+
"loss": 0.2319,
|
| 1816 |
+
"step": 258
|
| 1817 |
+
},
|
| 1818 |
+
{
|
| 1819 |
+
"epoch": 1.1671372106154716,
|
| 1820 |
+
"grad_norm": 1.3099559545516968,
|
| 1821 |
+
"learning_rate": 4.46952780188471e-06,
|
| 1822 |
+
"loss": 0.2305,
|
| 1823 |
+
"step": 259
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"epoch": 1.1716544325239977,
|
| 1827 |
+
"grad_norm": 1.16286039352417,
|
| 1828 |
+
"learning_rate": 4.4303985130598615e-06,
|
| 1829 |
+
"loss": 0.2231,
|
| 1830 |
+
"step": 260
|
| 1831 |
+
},
|
| 1832 |
+
{
|
| 1833 |
+
"epoch": 1.176171654432524,
|
| 1834 |
+
"grad_norm": 1.0713131427764893,
|
| 1835 |
+
"learning_rate": 4.391304536276389e-06,
|
| 1836 |
+
"loss": 0.2132,
|
| 1837 |
+
"step": 261
|
| 1838 |
+
},
|
| 1839 |
+
{
|
| 1840 |
+
"epoch": 1.1806888763410504,
|
| 1841 |
+
"grad_norm": 1.2314507961273193,
|
| 1842 |
+
"learning_rate": 4.352248295137739e-06,
|
| 1843 |
+
"loss": 0.2181,
|
| 1844 |
+
"step": 262
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 1.1852060982495765,
|
| 1848 |
+
"grad_norm": 1.4253009557724,
|
| 1849 |
+
"learning_rate": 4.313232210907959e-06,
|
| 1850 |
+
"loss": 0.2205,
|
| 1851 |
+
"step": 263
|
| 1852 |
+
},
|
| 1853 |
+
{
|
| 1854 |
+
"epoch": 1.1897233201581028,
|
| 1855 |
+
"grad_norm": 1.146643042564392,
|
| 1856 |
+
"learning_rate": 4.274258702361604e-06,
|
| 1857 |
+
"loss": 0.2192,
|
| 1858 |
+
"step": 264
|
| 1859 |
+
},
|
| 1860 |
+
{
|
| 1861 |
+
"epoch": 1.194240542066629,
|
| 1862 |
+
"grad_norm": 0.9879461526870728,
|
| 1863 |
+
"learning_rate": 4.23533018563377e-06,
|
| 1864 |
+
"loss": 0.2155,
|
| 1865 |
+
"step": 265
|
| 1866 |
+
},
|
| 1867 |
+
{
|
| 1868 |
+
"epoch": 1.1987577639751552,
|
| 1869 |
+
"grad_norm": 1.0763574838638306,
|
| 1870 |
+
"learning_rate": 4.196449074070329e-06,
|
| 1871 |
+
"loss": 0.2136,
|
| 1872 |
+
"step": 266
|
| 1873 |
+
},
|
| 1874 |
+
{
|
| 1875 |
+
"epoch": 1.2032749858836815,
|
| 1876 |
+
"grad_norm": 1.3163200616836548,
|
| 1877 |
+
"learning_rate": 4.15761777807829e-06,
|
| 1878 |
+
"loss": 0.2128,
|
| 1879 |
+
"step": 267
|
| 1880 |
+
},
|
| 1881 |
+
{
|
| 1882 |
+
"epoch": 1.2077922077922079,
|
| 1883 |
+
"grad_norm": 1.3158272504806519,
|
| 1884 |
+
"learning_rate": 4.118838704976392e-06,
|
| 1885 |
+
"loss": 0.213,
|
| 1886 |
+
"step": 268
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 1.212309429700734,
|
| 1890 |
+
"grad_norm": 1.124817132949829,
|
| 1891 |
+
"learning_rate": 4.080114258845846e-06,
|
| 1892 |
+
"loss": 0.2186,
|
| 1893 |
+
"step": 269
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 1.2168266516092603,
|
| 1897 |
+
"grad_norm": 1.1138076782226562,
|
| 1898 |
+
"learning_rate": 4.041446840381309e-06,
|
| 1899 |
+
"loss": 0.2208,
|
| 1900 |
+
"step": 270
|
| 1901 |
+
},
|
| 1902 |
+
{
|
| 1903 |
+
"epoch": 1.2213438735177866,
|
| 1904 |
+
"grad_norm": 1.3045446872711182,
|
| 1905 |
+
"learning_rate": 4.002838846742039e-06,
|
| 1906 |
+
"loss": 0.2183,
|
| 1907 |
+
"step": 271
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"epoch": 1.2258610954263127,
|
| 1911 |
+
"grad_norm": 1.0799813270568848,
|
| 1912 |
+
"learning_rate": 3.964292671403303e-06,
|
| 1913 |
+
"loss": 0.212,
|
| 1914 |
+
"step": 272
|
| 1915 |
+
},
|
| 1916 |
+
{
|
| 1917 |
+
"epoch": 1.230378317334839,
|
| 1918 |
+
"grad_norm": 1.0639320611953735,
|
| 1919 |
+
"learning_rate": 3.92581070400798e-06,
|
| 1920 |
+
"loss": 0.2146,
|
| 1921 |
+
"step": 273
|
| 1922 |
+
},
|
| 1923 |
+
{
|
| 1924 |
+
"epoch": 1.2348955392433654,
|
| 1925 |
+
"grad_norm": 1.0948545932769775,
|
| 1926 |
+
"learning_rate": 3.887395330218429e-06,
|
| 1927 |
+
"loss": 0.2202,
|
| 1928 |
+
"step": 274
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 1.2394127611518915,
|
| 1932 |
+
"grad_norm": 1.1420738697052002,
|
| 1933 |
+
"learning_rate": 3.8490489315685764e-06,
|
| 1934 |
+
"loss": 0.2165,
|
| 1935 |
+
"step": 275
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 1.2439299830604178,
|
| 1939 |
+
"grad_norm": 1.0136562585830688,
|
| 1940 |
+
"learning_rate": 3.8107738853162953e-06,
|
| 1941 |
+
"loss": 0.2217,
|
| 1942 |
+
"step": 276
|
| 1943 |
+
},
|
| 1944 |
+
{
|
| 1945 |
+
"epoch": 1.2484472049689441,
|
| 1946 |
+
"grad_norm": 1.0851843357086182,
|
| 1947 |
+
"learning_rate": 3.7725725642960047e-06,
|
| 1948 |
+
"loss": 0.2156,
|
| 1949 |
+
"step": 277
|
| 1950 |
+
},
|
| 1951 |
+
{
|
| 1952 |
+
"epoch": 1.2529644268774702,
|
| 1953 |
+
"grad_norm": 0.965035617351532,
|
| 1954 |
+
"learning_rate": 3.734447336771591e-06,
|
| 1955 |
+
"loss": 0.2169,
|
| 1956 |
+
"step": 278
|
| 1957 |
+
},
|
| 1958 |
+
{
|
| 1959 |
+
"epoch": 1.2574816487859966,
|
| 1960 |
+
"grad_norm": 0.9297024011611938,
|
| 1961 |
+
"learning_rate": 3.696400566289571e-06,
|
| 1962 |
+
"loss": 0.2265,
|
| 1963 |
+
"step": 279
|
| 1964 |
+
},
|
| 1965 |
+
{
|
| 1966 |
+
"epoch": 1.2619988706945229,
|
| 1967 |
+
"grad_norm": 0.9791793823242188,
|
| 1968 |
+
"learning_rate": 3.658434611532578e-06,
|
| 1969 |
+
"loss": 0.2223,
|
| 1970 |
+
"step": 280
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 1.2665160926030492,
|
| 1974 |
+
"grad_norm": 1.024862289428711,
|
| 1975 |
+
"learning_rate": 3.6205518261731247e-06,
|
| 1976 |
+
"loss": 0.2141,
|
| 1977 |
+
"step": 281
|
| 1978 |
+
},
|
| 1979 |
+
{
|
| 1980 |
+
"epoch": 1.2710333145115753,
|
| 1981 |
+
"grad_norm": 1.0613692998886108,
|
| 1982 |
+
"learning_rate": 3.5827545587277033e-06,
|
| 1983 |
+
"loss": 0.2175,
|
| 1984 |
+
"step": 282
|
| 1985 |
+
},
|
| 1986 |
+
{
|
| 1987 |
+
"epoch": 1.2755505364201016,
|
| 1988 |
+
"grad_norm": 1.042018175125122,
|
| 1989 |
+
"learning_rate": 3.545045152411178e-06,
|
| 1990 |
+
"loss": 0.2163,
|
| 1991 |
+
"step": 283
|
| 1992 |
+
},
|
| 1993 |
+
{
|
| 1994 |
+
"epoch": 1.280067758328628,
|
| 1995 |
+
"grad_norm": 0.9863473773002625,
|
| 1996 |
+
"learning_rate": 3.507425944991529e-06,
|
| 1997 |
+
"loss": 0.215,
|
| 1998 |
+
"step": 284
|
| 1999 |
+
},
|
| 2000 |
+
{
|
| 2001 |
+
"epoch": 1.2845849802371543,
|
| 2002 |
+
"grad_norm": 1.04150390625,
|
| 2003 |
+
"learning_rate": 3.469899268644913e-06,
|
| 2004 |
+
"loss": 0.219,
|
| 2005 |
+
"step": 285
|
| 2006 |
+
},
|
| 2007 |
+
{
|
| 2008 |
+
"epoch": 1.2891022021456804,
|
| 2009 |
+
"grad_norm": 1.1645381450653076,
|
| 2010 |
+
"learning_rate": 3.4324674498110956e-06,
|
| 2011 |
+
"loss": 0.2185,
|
| 2012 |
+
"step": 286
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 1.2936194240542067,
|
| 2016 |
+
"grad_norm": 1.0116959810256958,
|
| 2017 |
+
"learning_rate": 3.395132809049212e-06,
|
| 2018 |
+
"loss": 0.2244,
|
| 2019 |
+
"step": 287
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 1.298136645962733,
|
| 2023 |
+
"grad_norm": 1.063246726989746,
|
| 2024 |
+
"learning_rate": 3.3578976608939184e-06,
|
| 2025 |
+
"loss": 0.2228,
|
| 2026 |
+
"step": 288
|
| 2027 |
+
},
|
| 2028 |
+
{
|
| 2029 |
+
"epoch": 1.3026538678712591,
|
| 2030 |
+
"grad_norm": 1.0815736055374146,
|
| 2031 |
+
"learning_rate": 3.3207643137118872e-06,
|
| 2032 |
+
"loss": 0.2181,
|
| 2033 |
+
"step": 289
|
| 2034 |
+
},
|
| 2035 |
+
{
|
| 2036 |
+
"epoch": 1.3071710897797855,
|
| 2037 |
+
"grad_norm": 1.100237488746643,
|
| 2038 |
+
"learning_rate": 3.2837350695587237e-06,
|
| 2039 |
+
"loss": 0.2192,
|
| 2040 |
+
"step": 290
|
| 2041 |
+
},
|
| 2042 |
+
{
|
| 2043 |
+
"epoch": 1.3116883116883118,
|
| 2044 |
+
"grad_norm": 1.035388469696045,
|
| 2045 |
+
"learning_rate": 3.2468122240362287e-06,
|
| 2046 |
+
"loss": 0.233,
|
| 2047 |
+
"step": 291
|
| 2048 |
+
},
|
| 2049 |
+
{
|
| 2050 |
+
"epoch": 1.316205533596838,
|
| 2051 |
+
"grad_norm": 0.9863721132278442,
|
| 2052 |
+
"learning_rate": 3.2099980661501016e-06,
|
| 2053 |
+
"loss": 0.2135,
|
| 2054 |
+
"step": 292
|
| 2055 |
+
},
|
| 2056 |
+
{
|
| 2057 |
+
"epoch": 1.3207227555053642,
|
| 2058 |
+
"grad_norm": 1.0871480703353882,
|
| 2059 |
+
"learning_rate": 3.173294878168025e-06,
|
| 2060 |
+
"loss": 0.2035,
|
| 2061 |
+
"step": 293
|
| 2062 |
+
},
|
| 2063 |
+
{
|
| 2064 |
+
"epoch": 1.3252399774138905,
|
| 2065 |
+
"grad_norm": 1.0046381950378418,
|
| 2066 |
+
"learning_rate": 3.1367049354781854e-06,
|
| 2067 |
+
"loss": 0.2086,
|
| 2068 |
+
"step": 294
|
| 2069 |
+
},
|
| 2070 |
+
{
|
| 2071 |
+
"epoch": 1.3297571993224166,
|
| 2072 |
+
"grad_norm": 0.9969596862792969,
|
| 2073 |
+
"learning_rate": 3.1002305064482006e-06,
|
| 2074 |
+
"loss": 0.2161,
|
| 2075 |
+
"step": 295
|
| 2076 |
+
},
|
| 2077 |
+
{
|
| 2078 |
+
"epoch": 1.334274421230943,
|
| 2079 |
+
"grad_norm": 0.9493130445480347,
|
| 2080 |
+
"learning_rate": 3.063873852284508e-06,
|
| 2081 |
+
"loss": 0.2221,
|
| 2082 |
+
"step": 296
|
| 2083 |
+
},
|
| 2084 |
+
{
|
| 2085 |
+
"epoch": 1.3387916431394693,
|
| 2086 |
+
"grad_norm": 0.9661362171173096,
|
| 2087 |
+
"learning_rate": 3.0276372268921694e-06,
|
| 2088 |
+
"loss": 0.215,
|
| 2089 |
+
"step": 297
|
| 2090 |
+
},
|
| 2091 |
+
{
|
| 2092 |
+
"epoch": 1.3433088650479954,
|
| 2093 |
+
"grad_norm": 0.9412975907325745,
|
| 2094 |
+
"learning_rate": 2.991522876735154e-06,
|
| 2095 |
+
"loss": 0.2115,
|
| 2096 |
+
"step": 298
|
| 2097 |
+
},
|
| 2098 |
+
{
|
| 2099 |
+
"epoch": 1.3478260869565217,
|
| 2100 |
+
"grad_norm": 1.0934168100357056,
|
| 2101 |
+
"learning_rate": 2.9555330406970568e-06,
|
| 2102 |
+
"loss": 0.2172,
|
| 2103 |
+
"step": 299
|
| 2104 |
+
},
|
| 2105 |
+
{
|
| 2106 |
+
"epoch": 1.352343308865048,
|
| 2107 |
+
"grad_norm": 0.9448551535606384,
|
| 2108 |
+
"learning_rate": 2.9196699499423143e-06,
|
| 2109 |
+
"loss": 0.1957,
|
| 2110 |
+
"step": 300
|
| 2111 |
+
},
|
| 2112 |
+
{
|
| 2113 |
+
"epoch": 1.3568605307735742,
|
| 2114 |
+
"grad_norm": 1.0115864276885986,
|
| 2115 |
+
"learning_rate": 2.8839358277778758e-06,
|
| 2116 |
+
"loss": 0.2181,
|
| 2117 |
+
"step": 301
|
| 2118 |
+
},
|
| 2119 |
+
{
|
| 2120 |
+
"epoch": 1.3613777526821005,
|
| 2121 |
+
"grad_norm": 1.00177800655365,
|
| 2122 |
+
"learning_rate": 2.848332889515375e-06,
|
| 2123 |
+
"loss": 0.2078,
|
| 2124 |
+
"step": 302
|
| 2125 |
+
},
|
| 2126 |
+
{
|
| 2127 |
+
"epoch": 1.3658949745906268,
|
| 2128 |
+
"grad_norm": 1.0666226148605347,
|
| 2129 |
+
"learning_rate": 2.8128633423337932e-06,
|
| 2130 |
+
"loss": 0.215,
|
| 2131 |
+
"step": 303
|
| 2132 |
+
},
|
| 2133 |
+
{
|
| 2134 |
+
"epoch": 1.370412196499153,
|
| 2135 |
+
"grad_norm": 1.0075701475143433,
|
| 2136 |
+
"learning_rate": 2.7775293851426233e-06,
|
| 2137 |
+
"loss": 0.2151,
|
| 2138 |
+
"step": 304
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 1.3749294184076792,
|
| 2142 |
+
"grad_norm": 0.9858704209327698,
|
| 2143 |
+
"learning_rate": 2.7423332084455543e-06,
|
| 2144 |
+
"loss": 0.2166,
|
| 2145 |
+
"step": 305
|
| 2146 |
+
},
|
| 2147 |
+
{
|
| 2148 |
+
"epoch": 1.3794466403162056,
|
| 2149 |
+
"grad_norm": 0.9559859037399292,
|
| 2150 |
+
"learning_rate": 2.7072769942046716e-06,
|
| 2151 |
+
"loss": 0.1981,
|
| 2152 |
+
"step": 306
|
| 2153 |
+
},
|
| 2154 |
+
{
|
| 2155 |
+
"epoch": 1.3839638622247317,
|
| 2156 |
+
"grad_norm": 0.9518159031867981,
|
| 2157 |
+
"learning_rate": 2.6723629157051844e-06,
|
| 2158 |
+
"loss": 0.2069,
|
| 2159 |
+
"step": 307
|
| 2160 |
+
},
|
| 2161 |
+
{
|
| 2162 |
+
"epoch": 1.388481084133258,
|
| 2163 |
+
"grad_norm": 0.9317282438278198,
|
| 2164 |
+
"learning_rate": 2.637593137420702e-06,
|
| 2165 |
+
"loss": 0.2136,
|
| 2166 |
+
"step": 308
|
| 2167 |
+
},
|
| 2168 |
+
{
|
| 2169 |
+
"epoch": 1.3929983060417843,
|
| 2170 |
+
"grad_norm": 1.0143823623657227,
|
| 2171 |
+
"learning_rate": 2.6029698148790392e-06,
|
| 2172 |
+
"loss": 0.2172,
|
| 2173 |
+
"step": 309
|
| 2174 |
+
},
|
| 2175 |
+
{
|
| 2176 |
+
"epoch": 1.3975155279503104,
|
| 2177 |
+
"grad_norm": 1.065935730934143,
|
| 2178 |
+
"learning_rate": 2.5684950945285937e-06,
|
| 2179 |
+
"loss": 0.2193,
|
| 2180 |
+
"step": 310
|
| 2181 |
+
},
|
| 2182 |
+
{
|
| 2183 |
+
"epoch": 1.4020327498588367,
|
| 2184 |
+
"grad_norm": 1.057476282119751,
|
| 2185 |
+
"learning_rate": 2.5341711136052728e-06,
|
| 2186 |
+
"loss": 0.2126,
|
| 2187 |
+
"step": 311
|
| 2188 |
+
},
|
| 2189 |
+
{
|
| 2190 |
+
"epoch": 1.406549971767363,
|
| 2191 |
+
"grad_norm": 0.9720900058746338,
|
| 2192 |
+
"learning_rate": 2.5000000000000015e-06,
|
| 2193 |
+
"loss": 0.2255,
|
| 2194 |
+
"step": 312
|
| 2195 |
+
},
|
| 2196 |
+
{
|
| 2197 |
+
"epoch": 1.4110671936758894,
|
| 2198 |
+
"grad_norm": 1.0138583183288574,
|
| 2199 |
+
"learning_rate": 2.4659838721268005e-06,
|
| 2200 |
+
"loss": 0.2025,
|
| 2201 |
+
"step": 313
|
| 2202 |
+
},
|
| 2203 |
+
{
|
| 2204 |
+
"epoch": 1.4155844155844157,
|
| 2205 |
+
"grad_norm": 1.0183136463165283,
|
| 2206 |
+
"learning_rate": 2.4321248387914677e-06,
|
| 2207 |
+
"loss": 0.2118,
|
| 2208 |
+
"step": 314
|
| 2209 |
+
},
|
| 2210 |
+
{
|
| 2211 |
+
"epoch": 1.4201016374929418,
|
| 2212 |
+
"grad_norm": 1.047669529914856,
|
| 2213 |
+
"learning_rate": 2.3984249990608237e-06,
|
| 2214 |
+
"loss": 0.2075,
|
| 2215 |
+
"step": 315
|
| 2216 |
+
},
|
| 2217 |
+
{
|
| 2218 |
+
"epoch": 1.4246188594014682,
|
| 2219 |
+
"grad_norm": 1.0808067321777344,
|
| 2220 |
+
"learning_rate": 2.364886442132606e-06,
|
| 2221 |
+
"loss": 0.2201,
|
| 2222 |
+
"step": 316
|
| 2223 |
+
},
|
| 2224 |
+
{
|
| 2225 |
+
"epoch": 1.4291360813099945,
|
| 2226 |
+
"grad_norm": 1.0421336889266968,
|
| 2227 |
+
"learning_rate": 2.331511247205933e-06,
|
| 2228 |
+
"loss": 0.2113,
|
| 2229 |
+
"step": 317
|
| 2230 |
+
},
|
| 2231 |
+
{
|
| 2232 |
+
"epoch": 1.4336533032185206,
|
| 2233 |
+
"grad_norm": 0.9820194840431213,
|
| 2234 |
+
"learning_rate": 2.2983014833524115e-06,
|
| 2235 |
+
"loss": 0.2126,
|
| 2236 |
+
"step": 318
|
| 2237 |
+
},
|
| 2238 |
+
{
|
| 2239 |
+
"epoch": 1.438170525127047,
|
| 2240 |
+
"grad_norm": 1.0598773956298828,
|
| 2241 |
+
"learning_rate": 2.265259209387867e-06,
|
| 2242 |
+
"loss": 0.2106,
|
| 2243 |
+
"step": 319
|
| 2244 |
+
},
|
| 2245 |
+
{
|
| 2246 |
+
"epoch": 1.4426877470355732,
|
| 2247 |
+
"grad_norm": 1.003958821296692,
|
| 2248 |
+
"learning_rate": 2.2323864737447067e-06,
|
| 2249 |
+
"loss": 0.2055,
|
| 2250 |
+
"step": 320
|
| 2251 |
+
},
|
| 2252 |
+
{
|
| 2253 |
+
"epoch": 1.4472049689440993,
|
| 2254 |
+
"grad_norm": 1.071560263633728,
|
| 2255 |
+
"learning_rate": 2.1996853143449285e-06,
|
| 2256 |
+
"loss": 0.2094,
|
| 2257 |
+
"step": 321
|
| 2258 |
+
},
|
| 2259 |
+
{
|
| 2260 |
+
"epoch": 1.4517221908526257,
|
| 2261 |
+
"grad_norm": 1.0701788663864136,
|
| 2262 |
+
"learning_rate": 2.16715775847379e-06,
|
| 2263 |
+
"loss": 0.215,
|
| 2264 |
+
"step": 322
|
| 2265 |
+
},
|
| 2266 |
+
{
|
| 2267 |
+
"epoch": 1.456239412761152,
|
| 2268 |
+
"grad_norm": 1.074393391609192,
|
| 2269 |
+
"learning_rate": 2.1348058226541072e-06,
|
| 2270 |
+
"loss": 0.219,
|
| 2271 |
+
"step": 323
|
| 2272 |
+
},
|
| 2273 |
+
{
|
| 2274 |
+
"epoch": 1.460756634669678,
|
| 2275 |
+
"grad_norm": 1.0116572380065918,
|
| 2276 |
+
"learning_rate": 2.102631512521269e-06,
|
| 2277 |
+
"loss": 0.2034,
|
| 2278 |
+
"step": 324
|
| 2279 |
+
},
|
| 2280 |
+
{
|
| 2281 |
+
"epoch": 1.4652738565782044,
|
| 2282 |
+
"grad_norm": 0.9663606882095337,
|
| 2283 |
+
"learning_rate": 2.0706368226988772e-06,
|
| 2284 |
+
"loss": 0.2083,
|
| 2285 |
+
"step": 325
|
| 2286 |
+
},
|
| 2287 |
+
{
|
| 2288 |
+
"epoch": 1.4697910784867307,
|
| 2289 |
+
"grad_norm": 1.075634479522705,
|
| 2290 |
+
"learning_rate": 2.0388237366751005e-06,
|
| 2291 |
+
"loss": 0.2083,
|
| 2292 |
+
"step": 326
|
| 2293 |
+
},
|
| 2294 |
+
{
|
| 2295 |
+
"epoch": 1.4743083003952568,
|
| 2296 |
+
"grad_norm": 0.9744548201560974,
|
| 2297 |
+
"learning_rate": 2.00719422667971e-06,
|
| 2298 |
+
"loss": 0.2184,
|
| 2299 |
+
"step": 327
|
| 2300 |
+
},
|
| 2301 |
+
{
|
| 2302 |
+
"epoch": 1.4788255223037832,
|
| 2303 |
+
"grad_norm": 1.0275142192840576,
|
| 2304 |
+
"learning_rate": 1.9757502535618137e-06,
|
| 2305 |
+
"loss": 0.1991,
|
| 2306 |
+
"step": 328
|
| 2307 |
+
},
|
| 2308 |
+
{
|
| 2309 |
+
"epoch": 1.4833427442123095,
|
| 2310 |
+
"grad_norm": 0.950013279914856,
|
| 2311 |
+
"learning_rate": 1.9444937666682834e-06,
|
| 2312 |
+
"loss": 0.2082,
|
| 2313 |
+
"step": 329
|
| 2314 |
+
},
|
| 2315 |
+
{
|
| 2316 |
+
"epoch": 1.4878599661208356,
|
| 2317 |
+
"grad_norm": 0.9861937165260315,
|
| 2318 |
+
"learning_rate": 1.913426703722924e-06,
|
| 2319 |
+
"loss": 0.2217,
|
| 2320 |
+
"step": 330
|
| 2321 |
+
},
|
| 2322 |
+
{
|
| 2323 |
+
"epoch": 1.492377188029362,
|
| 2324 |
+
"grad_norm": 1.1520744562149048,
|
| 2325 |
+
"learning_rate": 1.8825509907063328e-06,
|
| 2326 |
+
"loss": 0.2107,
|
| 2327 |
+
"step": 331
|
| 2328 |
+
},
|
| 2329 |
+
{
|
| 2330 |
+
"epoch": 1.4968944099378882,
|
| 2331 |
+
"grad_norm": 1.03274667263031,
|
| 2332 |
+
"learning_rate": 1.851868541736503e-06,
|
| 2333 |
+
"loss": 0.2039,
|
| 2334 |
+
"step": 332
|
| 2335 |
+
},
|
| 2336 |
+
{
|
| 2337 |
+
"epoch": 1.5014116318464144,
|
| 2338 |
+
"grad_norm": 0.949181079864502,
|
| 2339 |
+
"learning_rate": 1.8213812589501611e-06,
|
| 2340 |
+
"loss": 0.2142,
|
| 2341 |
+
"step": 333
|
| 2342 |
+
},
|
| 2343 |
+
{
|
| 2344 |
+
"epoch": 1.5059288537549407,
|
| 2345 |
+
"grad_norm": 0.9323081374168396,
|
| 2346 |
+
"learning_rate": 1.7910910323848435e-06,
|
| 2347 |
+
"loss": 0.1976,
|
| 2348 |
+
"step": 334
|
| 2349 |
+
},
|
| 2350 |
+
{
|
| 2351 |
+
"epoch": 1.510446075663467,
|
| 2352 |
+
"grad_norm": 1.1117639541625977,
|
| 2353 |
+
"learning_rate": 1.760999739861724e-06,
|
| 2354 |
+
"loss": 0.2126,
|
| 2355 |
+
"step": 335
|
| 2356 |
+
},
|
| 2357 |
+
{
|
| 2358 |
+
"epoch": 1.514963297571993,
|
| 2359 |
+
"grad_norm": 1.0425753593444824,
|
| 2360 |
+
"learning_rate": 1.7311092468692082e-06,
|
| 2361 |
+
"loss": 0.2178,
|
| 2362 |
+
"step": 336
|
| 2363 |
+
},
|
| 2364 |
+
{
|
| 2365 |
+
"epoch": 1.5194805194805194,
|
| 2366 |
+
"grad_norm": 0.9980480074882507,
|
| 2367 |
+
"learning_rate": 1.7014214064472646e-06,
|
| 2368 |
+
"loss": 0.205,
|
| 2369 |
+
"step": 337
|
| 2370 |
+
},
|
| 2371 |
+
{
|
| 2372 |
+
"epoch": 1.5239977413890458,
|
| 2373 |
+
"grad_norm": 0.901735782623291,
|
| 2374 |
+
"learning_rate": 1.671938059072571e-06,
|
| 2375 |
+
"loss": 0.1996,
|
| 2376 |
+
"step": 338
|
| 2377 |
+
},
|
| 2378 |
+
{
|
| 2379 |
+
"epoch": 1.5285149632975719,
|
| 2380 |
+
"grad_norm": 1.0576623678207397,
|
| 2381 |
+
"learning_rate": 1.642661032544396e-06,
|
| 2382 |
+
"loss": 0.2024,
|
| 2383 |
+
"step": 339
|
| 2384 |
+
},
|
| 2385 |
+
{
|
| 2386 |
+
"epoch": 1.5330321852060984,
|
| 2387 |
+
"grad_norm": 1.1729637384414673,
|
| 2388 |
+
"learning_rate": 1.6135921418712959e-06,
|
| 2389 |
+
"loss": 0.2095,
|
| 2390 |
+
"step": 340
|
| 2391 |
+
},
|
| 2392 |
+
{
|
| 2393 |
+
"epoch": 1.5375494071146245,
|
| 2394 |
+
"grad_norm": 0.9823737144470215,
|
| 2395 |
+
"learning_rate": 1.5847331891585888e-06,
|
| 2396 |
+
"loss": 0.1995,
|
| 2397 |
+
"step": 341
|
| 2398 |
+
},
|
| 2399 |
+
{
|
| 2400 |
+
"epoch": 1.5420666290231506,
|
| 2401 |
+
"grad_norm": 0.9766678214073181,
|
| 2402 |
+
"learning_rate": 1.5560859634966457e-06,
|
| 2403 |
+
"loss": 0.2064,
|
| 2404 |
+
"step": 342
|
| 2405 |
+
},
|
| 2406 |
+
{
|
| 2407 |
+
"epoch": 1.5465838509316772,
|
| 2408 |
+
"grad_norm": 0.9916868209838867,
|
| 2409 |
+
"learning_rate": 1.5276522408499567e-06,
|
| 2410 |
+
"loss": 0.2164,
|
| 2411 |
+
"step": 343
|
| 2412 |
+
},
|
| 2413 |
+
{
|
| 2414 |
+
"epoch": 1.5511010728402033,
|
| 2415 |
+
"grad_norm": 0.9452782869338989,
|
| 2416 |
+
"learning_rate": 1.499433783947054e-06,
|
| 2417 |
+
"loss": 0.1971,
|
| 2418 |
+
"step": 344
|
| 2419 |
+
},
|
| 2420 |
+
{
|
| 2421 |
+
"epoch": 1.5556182947487294,
|
| 2422 |
+
"grad_norm": 0.9804942011833191,
|
| 2423 |
+
"learning_rate": 1.4714323421712163e-06,
|
| 2424 |
+
"loss": 0.2082,
|
| 2425 |
+
"step": 345
|
| 2426 |
+
},
|
| 2427 |
+
{
|
| 2428 |
+
"epoch": 1.560135516657256,
|
| 2429 |
+
"grad_norm": 0.9985517263412476,
|
| 2430 |
+
"learning_rate": 1.4436496514520253e-06,
|
| 2431 |
+
"loss": 0.2107,
|
| 2432 |
+
"step": 346
|
| 2433 |
+
},
|
| 2434 |
+
{
|
| 2435 |
+
"epoch": 1.564652738565782,
|
| 2436 |
+
"grad_norm": 0.9342421889305115,
|
| 2437 |
+
"learning_rate": 1.4160874341577447e-06,
|
| 2438 |
+
"loss": 0.214,
|
| 2439 |
+
"step": 347
|
| 2440 |
+
},
|
| 2441 |
+
{
|
| 2442 |
+
"epoch": 1.5691699604743083,
|
| 2443 |
+
"grad_norm": 1.0142865180969238,
|
| 2444 |
+
"learning_rate": 1.3887473989885441e-06,
|
| 2445 |
+
"loss": 0.2149,
|
| 2446 |
+
"step": 348
|
| 2447 |
+
},
|
| 2448 |
+
{
|
| 2449 |
+
"epoch": 1.5736871823828347,
|
| 2450 |
+
"grad_norm": 1.051569938659668,
|
| 2451 |
+
"learning_rate": 1.361631240870569e-06,
|
| 2452 |
+
"loss": 0.2027,
|
| 2453 |
+
"step": 349
|
| 2454 |
+
},
|
| 2455 |
+
{
|
| 2456 |
+
"epoch": 1.5782044042913608,
|
| 2457 |
+
"grad_norm": 0.9345071911811829,
|
| 2458 |
+
"learning_rate": 1.3347406408508695e-06,
|
| 2459 |
+
"loss": 0.2111,
|
| 2460 |
+
"step": 350
|
| 2461 |
+
},
|
| 2462 |
+
{
|
| 2463 |
+
"epoch": 1.582721626199887,
|
| 2464 |
+
"grad_norm": 0.9481232166290283,
|
| 2465 |
+
"learning_rate": 1.3080772659931728e-06,
|
| 2466 |
+
"loss": 0.2047,
|
| 2467 |
+
"step": 351
|
| 2468 |
+
},
|
| 2469 |
+
{
|
| 2470 |
+
"epoch": 1.5872388481084134,
|
| 2471 |
+
"grad_norm": 0.8910330533981323,
|
| 2472 |
+
"learning_rate": 1.281642769274552e-06,
|
| 2473 |
+
"loss": 0.2085,
|
| 2474 |
+
"step": 352
|
| 2475 |
+
},
|
| 2476 |
+
{
|
| 2477 |
+
"epoch": 1.5917560700169395,
|
| 2478 |
+
"grad_norm": 0.9882820248603821,
|
| 2479 |
+
"learning_rate": 1.255438789482935e-06,
|
| 2480 |
+
"loss": 0.2072,
|
| 2481 |
+
"step": 353
|
| 2482 |
+
},
|
| 2483 |
+
{
|
| 2484 |
+
"epoch": 1.5962732919254659,
|
| 2485 |
+
"grad_norm": 1.0289164781570435,
|
| 2486 |
+
"learning_rate": 1.2294669511155193e-06,
|
| 2487 |
+
"loss": 0.2148,
|
| 2488 |
+
"step": 354
|
| 2489 |
+
},
|
| 2490 |
+
{
|
| 2491 |
+
"epoch": 1.6007905138339922,
|
| 2492 |
+
"grad_norm": 1.0262577533721924,
|
| 2493 |
+
"learning_rate": 1.2037288642780575e-06,
|
| 2494 |
+
"loss": 0.2049,
|
| 2495 |
+
"step": 355
|
| 2496 |
+
},
|
| 2497 |
+
{
|
| 2498 |
+
"epoch": 1.6053077357425183,
|
| 2499 |
+
"grad_norm": 1.035154938697815,
|
| 2500 |
+
"learning_rate": 1.1782261245850417e-06,
|
| 2501 |
+
"loss": 0.2153,
|
| 2502 |
+
"step": 356
|
| 2503 |
+
},
|
| 2504 |
+
{
|
| 2505 |
+
"epoch": 1.6098249576510446,
|
| 2506 |
+
"grad_norm": 0.9348172545433044,
|
| 2507 |
+
"learning_rate": 1.1529603130607837e-06,
|
| 2508 |
+
"loss": 0.2179,
|
| 2509 |
+
"step": 357
|
| 2510 |
+
},
|
| 2511 |
+
{
|
| 2512 |
+
"epoch": 1.614342179559571,
|
| 2513 |
+
"grad_norm": 0.9265248775482178,
|
| 2514 |
+
"learning_rate": 1.1279329960414047e-06,
|
| 2515 |
+
"loss": 0.2087,
|
| 2516 |
+
"step": 358
|
| 2517 |
+
},
|
| 2518 |
+
{
|
| 2519 |
+
"epoch": 1.618859401468097,
|
| 2520 |
+
"grad_norm": 0.9629910588264465,
|
| 2521 |
+
"learning_rate": 1.1031457250777206e-06,
|
| 2522 |
+
"loss": 0.2153,
|
| 2523 |
+
"step": 359
|
| 2524 |
+
},
|
| 2525 |
+
{
|
| 2526 |
+
"epoch": 1.6233766233766234,
|
| 2527 |
+
"grad_norm": 1.017176866531372,
|
| 2528 |
+
"learning_rate": 1.0786000368390686e-06,
|
| 2529 |
+
"loss": 0.2122,
|
| 2530 |
+
"step": 360
|
| 2531 |
+
},
|
| 2532 |
+
{
|
| 2533 |
+
"epoch": 1.6278938452851497,
|
| 2534 |
+
"grad_norm": 1.0526764392852783,
|
| 2535 |
+
"learning_rate": 1.0542974530180327e-06,
|
| 2536 |
+
"loss": 0.2116,
|
| 2537 |
+
"step": 361
|
| 2538 |
+
},
|
| 2539 |
+
{
|
| 2540 |
+
"epoch": 1.6324110671936758,
|
| 2541 |
+
"grad_norm": 0.9504070281982422,
|
| 2542 |
+
"learning_rate": 1.0302394802361104e-06,
|
| 2543 |
+
"loss": 0.2245,
|
| 2544 |
+
"step": 362
|
| 2545 |
+
},
|
| 2546 |
+
{
|
| 2547 |
+
"epoch": 1.6369282891022021,
|
| 2548 |
+
"grad_norm": 0.9686842560768127,
|
| 2549 |
+
"learning_rate": 1.00642760995031e-06,
|
| 2550 |
+
"loss": 0.2034,
|
| 2551 |
+
"step": 363
|
| 2552 |
+
},
|
| 2553 |
+
{
|
| 2554 |
+
"epoch": 1.6414455110107284,
|
| 2555 |
+
"grad_norm": 1.1349575519561768,
|
| 2556 |
+
"learning_rate": 9.82863318360695e-07,
|
| 2557 |
+
"loss": 0.2122,
|
| 2558 |
+
"step": 364
|
| 2559 |
+
},
|
| 2560 |
+
{
|
| 2561 |
+
"epoch": 1.6459627329192545,
|
| 2562 |
+
"grad_norm": 0.9603061676025391,
|
| 2563 |
+
"learning_rate": 9.595480663188528e-07,
|
| 2564 |
+
"loss": 0.2034,
|
| 2565 |
+
"step": 365
|
| 2566 |
+
},
|
| 2567 |
+
{
|
| 2568 |
+
"epoch": 1.6504799548277809,
|
| 2569 |
+
"grad_norm": 0.8837600350379944,
|
| 2570 |
+
"learning_rate": 9.364832992373501e-07,
|
| 2571 |
+
"loss": 0.213,
|
| 2572 |
+
"step": 366
|
| 2573 |
+
},
|
| 2574 |
+
{
|
| 2575 |
+
"epoch": 1.6549971767363072,
|
| 2576 |
+
"grad_norm": 0.9571669101715088,
|
| 2577 |
+
"learning_rate": 9.136704470001101e-07,
|
| 2578 |
+
"loss": 0.2147,
|
| 2579 |
+
"step": 367
|
| 2580 |
+
},
|
| 2581 |
+
{
|
| 2582 |
+
"epoch": 1.6595143986448333,
|
| 2583 |
+
"grad_norm": 0.9975245594978333,
|
| 2584 |
+
"learning_rate": 8.911109238737748e-07,
|
| 2585 |
+
"loss": 0.2186,
|
| 2586 |
+
"step": 368
|
| 2587 |
+
},
|
| 2588 |
+
{
|
| 2589 |
+
"epoch": 1.6640316205533598,
|
| 2590 |
+
"grad_norm": 0.9019753336906433,
|
| 2591 |
+
"learning_rate": 8.688061284200266e-07,
|
| 2592 |
+
"loss": 0.1923,
|
| 2593 |
+
"step": 369
|
| 2594 |
+
},
|
| 2595 |
+
{
|
| 2596 |
+
"epoch": 1.668548842461886,
|
| 2597 |
+
"grad_norm": 0.9691225290298462,
|
| 2598 |
+
"learning_rate": 8.46757443408886e-07,
|
| 2599 |
+
"loss": 0.2056,
|
| 2600 |
+
"step": 370
|
| 2601 |
+
},
|
| 2602 |
+
{
|
| 2603 |
+
"epoch": 1.673066064370412,
|
| 2604 |
+
"grad_norm": 1.0318666696548462,
|
| 2605 |
+
"learning_rate": 8.24966235732988e-07,
|
| 2606 |
+
"loss": 0.215,
|
| 2607 |
+
"step": 371
|
| 2608 |
+
},
|
| 2609 |
+
{
|
| 2610 |
+
"epoch": 1.6775832862789386,
|
| 2611 |
+
"grad_norm": 0.8921061754226685,
|
| 2612 |
+
"learning_rate": 8.03433856322845e-07,
|
| 2613 |
+
"loss": 0.2176,
|
| 2614 |
+
"step": 372
|
| 2615 |
+
},
|
| 2616 |
+
{
|
| 2617 |
+
"epoch": 1.6821005081874647,
|
| 2618 |
+
"grad_norm": 0.982087254524231,
|
| 2619 |
+
"learning_rate": 7.821616400630866e-07,
|
| 2620 |
+
"loss": 0.2097,
|
| 2621 |
+
"step": 373
|
| 2622 |
+
},
|
| 2623 |
+
{
|
| 2624 |
+
"epoch": 1.6866177300959908,
|
| 2625 |
+
"grad_norm": 1.021616816520691,
|
| 2626 |
+
"learning_rate": 7.611509057097211e-07,
|
| 2627 |
+
"loss": 0.2156,
|
| 2628 |
+
"step": 374
|
| 2629 |
+
},
|
| 2630 |
+
{
|
| 2631 |
+
"epoch": 1.6911349520045174,
|
| 2632 |
+
"grad_norm": 1.0062682628631592,
|
| 2633 |
+
"learning_rate": 7.404029558083653e-07,
|
| 2634 |
+
"loss": 0.2139,
|
| 2635 |
+
"step": 375
|
| 2636 |
+
},
|
| 2637 |
+
{
|
| 2638 |
+
"epoch": 1.6956521739130435,
|
| 2639 |
+
"grad_norm": 1.0028045177459717,
|
| 2640 |
+
"learning_rate": 7.199190766135001e-07,
|
| 2641 |
+
"loss": 0.2086,
|
| 2642 |
+
"step": 376
|
| 2643 |
+
},
|
| 2644 |
+
{
|
| 2645 |
+
"epoch": 1.7001693958215698,
|
| 2646 |
+
"grad_norm": 0.9706637263298035,
|
| 2647 |
+
"learning_rate": 6.997005380087301e-07,
|
| 2648 |
+
"loss": 0.2235,
|
| 2649 |
+
"step": 377
|
| 2650 |
+
},
|
| 2651 |
+
{
|
| 2652 |
+
"epoch": 1.704686617730096,
|
| 2653 |
+
"grad_norm": 1.0096423625946045,
|
| 2654 |
+
"learning_rate": 6.797485934280618e-07,
|
| 2655 |
+
"loss": 0.204,
|
| 2656 |
+
"step": 378
|
| 2657 |
+
},
|
| 2658 |
+
{
|
| 2659 |
+
"epoch": 1.7092038396386222,
|
| 2660 |
+
"grad_norm": 1.1178041696548462,
|
| 2661 |
+
"learning_rate": 6.600644797781847e-07,
|
| 2662 |
+
"loss": 0.2172,
|
| 2663 |
+
"step": 379
|
| 2664 |
+
},
|
| 2665 |
+
{
|
| 2666 |
+
"epoch": 1.7137210615471485,
|
| 2667 |
+
"grad_norm": 0.9578860998153687,
|
| 2668 |
+
"learning_rate": 6.406494173618083e-07,
|
| 2669 |
+
"loss": 0.2014,
|
| 2670 |
+
"step": 380
|
| 2671 |
+
},
|
| 2672 |
+
{
|
| 2673 |
+
"epoch": 1.7182382834556749,
|
| 2674 |
+
"grad_norm": 0.9890934824943542,
|
| 2675 |
+
"learning_rate": 6.215046098019967e-07,
|
| 2676 |
+
"loss": 0.2057,
|
| 2677 |
+
"step": 381
|
| 2678 |
+
},
|
| 2679 |
+
{
|
| 2680 |
+
"epoch": 1.722755505364201,
|
| 2681 |
+
"grad_norm": 0.9775418639183044,
|
| 2682 |
+
"learning_rate": 6.026312439675553e-07,
|
| 2683 |
+
"loss": 0.2144,
|
| 2684 |
+
"step": 382
|
| 2685 |
+
},
|
| 2686 |
+
{
|
| 2687 |
+
"epoch": 1.7272727272727273,
|
| 2688 |
+
"grad_norm": 0.9456623792648315,
|
| 2689 |
+
"learning_rate": 5.84030489899452e-07,
|
| 2690 |
+
"loss": 0.205,
|
| 2691 |
+
"step": 383
|
| 2692 |
+
},
|
| 2693 |
+
{
|
| 2694 |
+
"epoch": 1.7317899491812536,
|
| 2695 |
+
"grad_norm": 0.9762837886810303,
|
| 2696 |
+
"learning_rate": 5.657035007382822e-07,
|
| 2697 |
+
"loss": 0.2107,
|
| 2698 |
+
"step": 384
|
| 2699 |
+
},
|
| 2700 |
+
{
|
| 2701 |
+
"epoch": 1.7363071710897797,
|
| 2702 |
+
"grad_norm": 0.896228551864624,
|
| 2703 |
+
"learning_rate": 5.476514126527771e-07,
|
| 2704 |
+
"loss": 0.2147,
|
| 2705 |
+
"step": 385
|
| 2706 |
+
},
|
| 2707 |
+
{
|
| 2708 |
+
"epoch": 1.740824392998306,
|
| 2709 |
+
"grad_norm": 0.9847128391265869,
|
| 2710 |
+
"learning_rate": 5.29875344769375e-07,
|
| 2711 |
+
"loss": 0.2195,
|
| 2712 |
+
"step": 386
|
| 2713 |
+
},
|
| 2714 |
+
{
|
| 2715 |
+
"epoch": 1.7453416149068324,
|
| 2716 |
+
"grad_norm": 0.9506767988204956,
|
| 2717 |
+
"learning_rate": 5.123763991028291e-07,
|
| 2718 |
+
"loss": 0.2047,
|
| 2719 |
+
"step": 387
|
| 2720 |
+
},
|
| 2721 |
+
{
|
| 2722 |
+
"epoch": 1.7498588368153585,
|
| 2723 |
+
"grad_norm": 0.8956678509712219,
|
| 2724 |
+
"learning_rate": 4.951556604879049e-07,
|
| 2725 |
+
"loss": 0.2087,
|
| 2726 |
+
"step": 388
|
| 2727 |
+
},
|
| 2728 |
+
{
|
| 2729 |
+
"epoch": 1.7543760587238848,
|
| 2730 |
+
"grad_norm": 0.9172845482826233,
|
| 2731 |
+
"learning_rate": 4.782141965121129e-07,
|
| 2732 |
+
"loss": 0.2142,
|
| 2733 |
+
"step": 389
|
| 2734 |
+
},
|
| 2735 |
+
{
|
| 2736 |
+
"epoch": 1.7588932806324111,
|
| 2737 |
+
"grad_norm": 0.9192934632301331,
|
| 2738 |
+
"learning_rate": 4.615530574495325e-07,
|
| 2739 |
+
"loss": 0.2109,
|
| 2740 |
+
"step": 390
|
| 2741 |
+
},
|
| 2742 |
+
{
|
| 2743 |
+
"epoch": 1.7634105025409372,
|
| 2744 |
+
"grad_norm": 0.9868291020393372,
|
| 2745 |
+
"learning_rate": 4.4517327619569784e-07,
|
| 2746 |
+
"loss": 0.2098,
|
| 2747 |
+
"step": 391
|
| 2748 |
+
},
|
| 2749 |
+
{
|
| 2750 |
+
"epoch": 1.7679277244494636,
|
| 2751 |
+
"grad_norm": 0.923726499080658,
|
| 2752 |
+
"learning_rate": 4.2907586820356337e-07,
|
| 2753 |
+
"loss": 0.2014,
|
| 2754 |
+
"step": 392
|
| 2755 |
+
},
|
| 2756 |
+
{
|
| 2757 |
+
"epoch": 1.7724449463579899,
|
| 2758 |
+
"grad_norm": 1.0496209859848022,
|
| 2759 |
+
"learning_rate": 4.132618314205544e-07,
|
| 2760 |
+
"loss": 0.2126,
|
| 2761 |
+
"step": 393
|
| 2762 |
+
},
|
| 2763 |
+
{
|
| 2764 |
+
"epoch": 1.776962168266516,
|
| 2765 |
+
"grad_norm": 0.9124986529350281,
|
| 2766 |
+
"learning_rate": 3.9773214622669974e-07,
|
| 2767 |
+
"loss": 0.2033,
|
| 2768 |
+
"step": 394
|
| 2769 |
+
},
|
| 2770 |
+
{
|
| 2771 |
+
"epoch": 1.7814793901750423,
|
| 2772 |
+
"grad_norm": 1.0255227088928223,
|
| 2773 |
+
"learning_rate": 3.8248777537384763e-07,
|
| 2774 |
+
"loss": 0.2057,
|
| 2775 |
+
"step": 395
|
| 2776 |
+
},
|
| 2777 |
+
{
|
| 2778 |
+
"epoch": 1.7859966120835686,
|
| 2779 |
+
"grad_norm": 0.9439749717712402,
|
| 2780 |
+
"learning_rate": 3.675296639259912e-07,
|
| 2781 |
+
"loss": 0.2103,
|
| 2782 |
+
"step": 396
|
| 2783 |
+
},
|
| 2784 |
+
{
|
| 2785 |
+
"epoch": 1.7905138339920947,
|
| 2786 |
+
"grad_norm": 0.9144439101219177,
|
| 2787 |
+
"learning_rate": 3.528587392006716e-07,
|
| 2788 |
+
"loss": 0.214,
|
| 2789 |
+
"step": 397
|
| 2790 |
+
},
|
| 2791 |
+
{
|
| 2792 |
+
"epoch": 1.795031055900621,
|
| 2793 |
+
"grad_norm": 0.8841158151626587,
|
| 2794 |
+
"learning_rate": 3.38475910711491e-07,
|
| 2795 |
+
"loss": 0.207,
|
| 2796 |
+
"step": 398
|
| 2797 |
+
},
|
| 2798 |
+
{
|
| 2799 |
+
"epoch": 1.7995482778091474,
|
| 2800 |
+
"grad_norm": 0.9291788935661316,
|
| 2801 |
+
"learning_rate": 3.243820701117306e-07,
|
| 2802 |
+
"loss": 0.2026,
|
| 2803 |
+
"step": 399
|
| 2804 |
+
},
|
| 2805 |
+
{
|
| 2806 |
+
"epoch": 1.8040654997176735,
|
| 2807 |
+
"grad_norm": 0.9121840000152588,
|
| 2808 |
+
"learning_rate": 3.105780911390738e-07,
|
| 2809 |
+
"loss": 0.2106,
|
| 2810 |
+
"step": 400
|
| 2811 |
+
},
|
| 2812 |
+
{
|
| 2813 |
+
"epoch": 1.8085827216262,
|
| 2814 |
+
"grad_norm": 0.9134534001350403,
|
| 2815 |
+
"learning_rate": 2.97064829561432e-07,
|
| 2816 |
+
"loss": 0.2014,
|
| 2817 |
+
"step": 401
|
| 2818 |
+
},
|
| 2819 |
+
{
|
| 2820 |
+
"epoch": 1.8130999435347261,
|
| 2821 |
+
"grad_norm": 0.9187113642692566,
|
| 2822 |
+
"learning_rate": 2.8384312312390306e-07,
|
| 2823 |
+
"loss": 0.212,
|
| 2824 |
+
"step": 402
|
| 2825 |
+
},
|
| 2826 |
+
{
|
| 2827 |
+
"epoch": 1.8176171654432522,
|
| 2828 |
+
"grad_norm": 1.0056486129760742,
|
| 2829 |
+
"learning_rate": 2.7091379149682683e-07,
|
| 2830 |
+
"loss": 0.2198,
|
| 2831 |
+
"step": 403
|
| 2832 |
+
},
|
| 2833 |
+
{
|
| 2834 |
+
"epoch": 1.8221343873517788,
|
| 2835 |
+
"grad_norm": 0.9583958387374878,
|
| 2836 |
+
"learning_rate": 2.582776362249739e-07,
|
| 2837 |
+
"loss": 0.2048,
|
| 2838 |
+
"step": 404
|
| 2839 |
+
},
|
| 2840 |
+
{
|
| 2841 |
+
"epoch": 1.826651609260305,
|
| 2842 |
+
"grad_norm": 0.9572198987007141,
|
| 2843 |
+
"learning_rate": 2.459354406778547e-07,
|
| 2844 |
+
"loss": 0.2075,
|
| 2845 |
+
"step": 405
|
| 2846 |
+
},
|
| 2847 |
+
{
|
| 2848 |
+
"epoch": 1.8311688311688312,
|
| 2849 |
+
"grad_norm": 0.911439061164856,
|
| 2850 |
+
"learning_rate": 2.3388797000115427e-07,
|
| 2851 |
+
"loss": 0.2068,
|
| 2852 |
+
"step": 406
|
| 2853 |
+
},
|
| 2854 |
+
{
|
| 2855 |
+
"epoch": 1.8356860530773575,
|
| 2856 |
+
"grad_norm": 0.8923501372337341,
|
| 2857 |
+
"learning_rate": 2.2213597106929608e-07,
|
| 2858 |
+
"loss": 0.2077,
|
| 2859 |
+
"step": 407
|
| 2860 |
+
},
|
| 2861 |
+
{
|
| 2862 |
+
"epoch": 1.8402032749858837,
|
| 2863 |
+
"grad_norm": 0.920656144618988,
|
| 2864 |
+
"learning_rate": 2.1068017243914663e-07,
|
| 2865 |
+
"loss": 0.1998,
|
| 2866 |
+
"step": 408
|
| 2867 |
+
},
|
| 2868 |
+
{
|
| 2869 |
+
"epoch": 1.84472049689441,
|
| 2870 |
+
"grad_norm": 0.9093790650367737,
|
| 2871 |
+
"learning_rate": 1.9952128430483718e-07,
|
| 2872 |
+
"loss": 0.2118,
|
| 2873 |
+
"step": 409
|
| 2874 |
+
},
|
| 2875 |
+
{
|
| 2876 |
+
"epoch": 1.8492377188029363,
|
| 2877 |
+
"grad_norm": 0.9370465278625488,
|
| 2878 |
+
"learning_rate": 1.8865999845374794e-07,
|
| 2879 |
+
"loss": 0.2041,
|
| 2880 |
+
"step": 410
|
| 2881 |
+
},
|
| 2882 |
+
{
|
| 2883 |
+
"epoch": 1.8537549407114624,
|
| 2884 |
+
"grad_norm": 0.9633747935295105,
|
| 2885 |
+
"learning_rate": 1.78096988223615e-07,
|
| 2886 |
+
"loss": 0.2107,
|
| 2887 |
+
"step": 411
|
| 2888 |
+
},
|
| 2889 |
+
{
|
| 2890 |
+
"epoch": 1.8582721626199887,
|
| 2891 |
+
"grad_norm": 0.8899579048156738,
|
| 2892 |
+
"learning_rate": 1.6783290846078714e-07,
|
| 2893 |
+
"loss": 0.2029,
|
| 2894 |
+
"step": 412
|
| 2895 |
+
},
|
| 2896 |
+
{
|
| 2897 |
+
"epoch": 1.862789384528515,
|
| 2898 |
+
"grad_norm": 1.003751516342163,
|
| 2899 |
+
"learning_rate": 1.5786839547963008e-07,
|
| 2900 |
+
"loss": 0.208,
|
| 2901 |
+
"step": 413
|
| 2902 |
+
},
|
| 2903 |
+
{
|
| 2904 |
+
"epoch": 1.8673066064370412,
|
| 2905 |
+
"grad_norm": 0.9010616540908813,
|
| 2906 |
+
"learning_rate": 1.4820406702308165e-07,
|
| 2907 |
+
"loss": 0.2075,
|
| 2908 |
+
"step": 414
|
| 2909 |
+
},
|
| 2910 |
+
{
|
| 2911 |
+
"epoch": 1.8718238283455675,
|
| 2912 |
+
"grad_norm": 0.8831467032432556,
|
| 2913 |
+
"learning_rate": 1.388405222243472e-07,
|
| 2914 |
+
"loss": 0.2034,
|
| 2915 |
+
"step": 415
|
| 2916 |
+
},
|
| 2917 |
+
{
|
| 2918 |
+
"epoch": 1.8763410502540938,
|
| 2919 |
+
"grad_norm": 0.8796132802963257,
|
| 2920 |
+
"learning_rate": 1.2977834156976733e-07,
|
| 2921 |
+
"loss": 0.2159,
|
| 2922 |
+
"step": 416
|
| 2923 |
+
},
|
| 2924 |
+
{
|
| 2925 |
+
"epoch": 1.88085827216262,
|
| 2926 |
+
"grad_norm": 0.9781243205070496,
|
| 2927 |
+
"learning_rate": 1.210180868628219e-07,
|
| 2928 |
+
"loss": 0.2078,
|
| 2929 |
+
"step": 417
|
| 2930 |
+
},
|
| 2931 |
+
{
|
| 2932 |
+
"epoch": 1.8853754940711462,
|
| 2933 |
+
"grad_norm": 0.942133903503418,
|
| 2934 |
+
"learning_rate": 1.1256030118930727e-07,
|
| 2935 |
+
"loss": 0.2048,
|
| 2936 |
+
"step": 418
|
| 2937 |
+
},
|
| 2938 |
+
{
|
| 2939 |
+
"epoch": 1.8898927159796726,
|
| 2940 |
+
"grad_norm": 0.9115908741950989,
|
| 2941 |
+
"learning_rate": 1.0440550888366485e-07,
|
| 2942 |
+
"loss": 0.2027,
|
| 2943 |
+
"step": 419
|
| 2944 |
+
},
|
| 2945 |
+
{
|
| 2946 |
+
"epoch": 1.8944099378881987,
|
| 2947 |
+
"grad_norm": 0.9324295520782471,
|
| 2948 |
+
"learning_rate": 9.655421549647603e-08,
|
| 2949 |
+
"loss": 0.2073,
|
| 2950 |
+
"step": 420
|
| 2951 |
+
},
|
| 2952 |
+
{
|
| 2953 |
+
"epoch": 1.898927159796725,
|
| 2954 |
+
"grad_norm": 0.9188907742500305,
|
| 2955 |
+
"learning_rate": 8.900690776312282e-08,
|
| 2956 |
+
"loss": 0.2054,
|
| 2957 |
+
"step": 421
|
| 2958 |
+
},
|
| 2959 |
+
{
|
| 2960 |
+
"epoch": 1.9034443817052513,
|
| 2961 |
+
"grad_norm": 0.8986204266548157,
|
| 2962 |
+
"learning_rate": 8.176405357361194e-08,
|
| 2963 |
+
"loss": 0.2162,
|
| 2964 |
+
"step": 422
|
| 2965 |
+
},
|
| 2966 |
+
{
|
| 2967 |
+
"epoch": 1.9079616036137774,
|
| 2968 |
+
"grad_norm": 0.8852443695068359,
|
| 2969 |
+
"learning_rate": 7.482610194356477e-08,
|
| 2970 |
+
"loss": 0.208,
|
| 2971 |
+
"step": 423
|
| 2972 |
+
},
|
| 2973 |
+
{
|
| 2974 |
+
"epoch": 1.9124788255223037,
|
| 2975 |
+
"grad_norm": 0.8807809948921204,
|
| 2976 |
+
"learning_rate": 6.819348298638839e-08,
|
| 2977 |
+
"loss": 0.1984,
|
| 2978 |
+
"step": 424
|
| 2979 |
+
},
|
| 2980 |
+
{
|
| 2981 |
+
"epoch": 1.91699604743083,
|
| 2982 |
+
"grad_norm": 0.9482453465461731,
|
| 2983 |
+
"learning_rate": 6.186660788660315e-08,
|
| 2984 |
+
"loss": 0.2058,
|
| 2985 |
+
"step": 425
|
| 2986 |
+
},
|
| 2987 |
+
{
|
| 2988 |
+
"epoch": 1.9215132693393562,
|
| 2989 |
+
"grad_norm": 0.891965389251709,
|
| 2990 |
+
"learning_rate": 5.584586887435739e-08,
|
| 2991 |
+
"loss": 0.2108,
|
| 2992 |
+
"step": 426
|
| 2993 |
+
},
|
| 2994 |
+
{
|
| 2995 |
+
"epoch": 1.9260304912478825,
|
| 2996 |
+
"grad_norm": 0.9519309997558594,
|
| 2997 |
+
"learning_rate": 5.013163920110864e-08,
|
| 2998 |
+
"loss": 0.2254,
|
| 2999 |
+
"step": 427
|
| 3000 |
+
},
|
| 3001 |
+
{
|
| 3002 |
+
"epoch": 1.9305477131564088,
|
| 3003 |
+
"grad_norm": 0.8655776381492615,
|
| 3004 |
+
"learning_rate": 4.4724273116483754e-08,
|
| 3005 |
+
"loss": 0.2061,
|
| 3006 |
+
"step": 428
|
| 3007 |
+
},
|
| 3008 |
+
{
|
| 3009 |
+
"epoch": 1.935064935064935,
|
| 3010 |
+
"grad_norm": 0.940295934677124,
|
| 3011 |
+
"learning_rate": 3.9624105846319813e-08,
|
| 3012 |
+
"loss": 0.2148,
|
| 3013 |
+
"step": 429
|
| 3014 |
+
},
|
| 3015 |
+
{
|
| 3016 |
+
"epoch": 1.9395821569734615,
|
| 3017 |
+
"grad_norm": 0.8930814862251282,
|
| 3018 |
+
"learning_rate": 3.483145357187967e-08,
|
| 3019 |
+
"loss": 0.2014,
|
| 3020 |
+
"step": 430
|
| 3021 |
+
},
|
| 3022 |
+
{
|
| 3023 |
+
"epoch": 1.9440993788819876,
|
| 3024 |
+
"grad_norm": 0.8650930523872375,
|
| 3025 |
+
"learning_rate": 3.034661341025258e-08,
|
| 3026 |
+
"loss": 0.2104,
|
| 3027 |
+
"step": 431
|
| 3028 |
+
},
|
| 3029 |
+
{
|
| 3030 |
+
"epoch": 1.9486166007905137,
|
| 3031 |
+
"grad_norm": 0.8346561789512634,
|
| 3032 |
+
"learning_rate": 2.6169863395932304e-08,
|
| 3033 |
+
"loss": 0.2003,
|
| 3034 |
+
"step": 432
|
| 3035 |
+
},
|
| 3036 |
+
{
|
| 3037 |
+
"epoch": 1.9531338226990402,
|
| 3038 |
+
"grad_norm": 0.9135526418685913,
|
| 3039 |
+
"learning_rate": 2.230146246358256e-08,
|
| 3040 |
+
"loss": 0.2174,
|
| 3041 |
+
"step": 433
|
| 3042 |
+
},
|
| 3043 |
+
{
|
| 3044 |
+
"epoch": 1.9576510446075663,
|
| 3045 |
+
"grad_norm": 0.9022271633148193,
|
| 3046 |
+
"learning_rate": 1.8741650431982615e-08,
|
| 3047 |
+
"loss": 0.2025,
|
| 3048 |
+
"step": 434
|
| 3049 |
+
},
|
| 3050 |
+
{
|
| 3051 |
+
"epoch": 1.9621682665160924,
|
| 3052 |
+
"grad_norm": 0.8826963901519775,
|
| 3053 |
+
"learning_rate": 1.549064798916311e-08,
|
| 3054 |
+
"loss": 0.2117,
|
| 3055 |
+
"step": 435
|
| 3056 |
+
},
|
| 3057 |
+
{
|
| 3058 |
+
"epoch": 1.966685488424619,
|
| 3059 |
+
"grad_norm": 0.9136354327201843,
|
| 3060 |
+
"learning_rate": 1.2548656678721404e-08,
|
| 3061 |
+
"loss": 0.2104,
|
| 3062 |
+
"step": 436
|
| 3063 |
+
},
|
| 3064 |
+
{
|
| 3065 |
+
"epoch": 1.971202710333145,
|
| 3066 |
+
"grad_norm": 0.9502382278442383,
|
| 3067 |
+
"learning_rate": 9.915858887327157e-09,
|
| 3068 |
+
"loss": 0.2189,
|
| 3069 |
+
"step": 437
|
| 3070 |
+
},
|
| 3071 |
+
{
|
| 3072 |
+
"epoch": 1.9757199322416714,
|
| 3073 |
+
"grad_norm": 0.9200468063354492,
|
| 3074 |
+
"learning_rate": 7.59241783341913e-09,
|
| 3075 |
+
"loss": 0.2026,
|
| 3076 |
+
"step": 438
|
| 3077 |
+
},
|
| 3078 |
+
{
|
| 3079 |
+
"epoch": 1.9802371541501977,
|
| 3080 |
+
"grad_norm": 0.8734921216964722,
|
| 3081 |
+
"learning_rate": 5.578477557081074e-09,
|
| 3082 |
+
"loss": 0.2134,
|
| 3083 |
+
"step": 439
|
| 3084 |
+
},
|
| 3085 |
+
{
|
| 3086 |
+
"epoch": 1.9847543760587238,
|
| 3087 |
+
"grad_norm": 0.9157832860946655,
|
| 3088 |
+
"learning_rate": 3.874162911117196e-09,
|
| 3089 |
+
"loss": 0.2081,
|
| 3090 |
+
"step": 440
|
| 3091 |
+
},
|
| 3092 |
+
{
|
| 3093 |
+
"epoch": 1.9892715979672502,
|
| 3094 |
+
"grad_norm": 0.8866028785705566,
|
| 3095 |
+
"learning_rate": 2.479579553307798e-09,
|
| 3096 |
+
"loss": 0.2027,
|
| 3097 |
+
"step": 441
|
| 3098 |
+
},
|
| 3099 |
+
{
|
| 3100 |
+
"epoch": 1.9937888198757765,
|
| 3101 |
+
"grad_norm": 0.9098318815231323,
|
| 3102 |
+
"learning_rate": 1.3948139398628492e-09,
|
| 3103 |
+
"loss": 0.1985,
|
| 3104 |
+
"step": 442
|
| 3105 |
+
},
|
| 3106 |
+
{
|
| 3107 |
+
"epoch": 1.9983060417843026,
|
| 3108 |
+
"grad_norm": 0.9027771949768066,
|
| 3109 |
+
"learning_rate": 6.199333200590519e-10,
|
| 3110 |
+
"loss": 0.1963,
|
| 3111 |
+
"step": 443
|
| 3112 |
+
},
|
| 3113 |
+
{
|
| 3114 |
+
"epoch": 2.0,
|
| 3115 |
+
"grad_norm": 1.5019222497940063,
|
| 3116 |
+
"learning_rate": 1.5498573207262112e-10,
|
| 3117 |
+
"loss": 0.2171,
|
| 3118 |
+
"step": 444
|
| 3119 |
+
},
|
| 3120 |
+
{
|
| 3121 |
+
"epoch": 2.0,
|
| 3122 |
+
"step": 444,
|
| 3123 |
+
"total_flos": 6.347064608565494e+17,
|
| 3124 |
+
"train_loss": 0.30420903603101634,
|
| 3125 |
+
"train_runtime": 1521.5027,
|
| 3126 |
+
"train_samples_per_second": 148.979,
|
| 3127 |
+
"train_steps_per_second": 0.292
|
| 3128 |
+
}
|
| 3129 |
+
],
|
| 3130 |
+
"logging_steps": 1.0,
|
| 3131 |
+
"max_steps": 444,
|
| 3132 |
+
"num_input_tokens_seen": 0,
|
| 3133 |
+
"num_train_epochs": 2,
|
| 3134 |
+
"save_steps": 500,
|
| 3135 |
+
"stateful_callbacks": {
|
| 3136 |
+
"TrainerControl": {
|
| 3137 |
+
"args": {
|
| 3138 |
+
"should_epoch_stop": false,
|
| 3139 |
+
"should_evaluate": false,
|
| 3140 |
+
"should_log": false,
|
| 3141 |
+
"should_save": true,
|
| 3142 |
+
"should_training_stop": true
|
| 3143 |
+
},
|
| 3144 |
+
"attributes": {}
|
| 3145 |
+
}
|
| 3146 |
+
},
|
| 3147 |
+
"total_flos": 6.347064608565494e+17,
|
| 3148 |
+
"train_batch_size": 8,
|
| 3149 |
+
"trial_name": null,
|
| 3150 |
+
"trial_params": null
|
| 3151 |
+
}
|
training_loss.png
ADDED
|