Model save
Browse files- README.md +76 -0
- all_results.json +9 -0
- train_results.json +9 -0
- trainer_state.json +2952 -0
README.md
ADDED
|
@@ -0,0 +1,76 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
library_name: peft
|
| 4 |
+
tags:
|
| 5 |
+
- trl
|
| 6 |
+
- sft
|
| 7 |
+
- generated_from_trainer
|
| 8 |
+
base_model: mistralai/Mistral-7B-Instruct-v0.2
|
| 9 |
+
model-index:
|
| 10 |
+
- name: Mistral-7B-Instruct-v0.2-miracl-raft-sft-v2.0
|
| 11 |
+
results: []
|
| 12 |
+
---
|
| 13 |
+
|
| 14 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 15 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 16 |
+
|
| 17 |
+
# Mistral-7B-Instruct-v0.2-miracl-raft-sft-v2.0
|
| 18 |
+
|
| 19 |
+
This model is a fine-tuned version of [mistralai/Mistral-7B-Instruct-v0.2](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2) on an unknown dataset.
|
| 20 |
+
It achieves the following results on the evaluation set:
|
| 21 |
+
- Loss: 1.2086
|
| 22 |
+
|
| 23 |
+
## Model description
|
| 24 |
+
|
| 25 |
+
More information needed
|
| 26 |
+
|
| 27 |
+
## Intended uses & limitations
|
| 28 |
+
|
| 29 |
+
More information needed
|
| 30 |
+
|
| 31 |
+
## Training and evaluation data
|
| 32 |
+
|
| 33 |
+
More information needed
|
| 34 |
+
|
| 35 |
+
## Training procedure
|
| 36 |
+
|
| 37 |
+
### Training hyperparameters
|
| 38 |
+
|
| 39 |
+
The following hyperparameters were used during training:
|
| 40 |
+
- learning_rate: 1e-05
|
| 41 |
+
- train_batch_size: 4
|
| 42 |
+
- eval_batch_size: 4
|
| 43 |
+
- seed: 42
|
| 44 |
+
- distributed_type: multi-GPU
|
| 45 |
+
- num_devices: 3
|
| 46 |
+
- gradient_accumulation_steps: 2
|
| 47 |
+
- total_train_batch_size: 24
|
| 48 |
+
- total_eval_batch_size: 12
|
| 49 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 50 |
+
- lr_scheduler_type: cosine
|
| 51 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 52 |
+
- num_epochs: 1
|
| 53 |
+
|
| 54 |
+
### Training results
|
| 55 |
+
|
| 56 |
+
| Training Loss | Epoch | Step | Validation Loss |
|
| 57 |
+
|:-------------:|:------:|:----:|:---------------:|
|
| 58 |
+
| 1.3095 | 0.0987 | 200 | 1.2800 |
|
| 59 |
+
| 1.249 | 0.1975 | 400 | 1.2516 |
|
| 60 |
+
| 1.2514 | 0.2962 | 600 | 1.2369 |
|
| 61 |
+
| 1.275 | 0.3950 | 800 | 1.2263 |
|
| 62 |
+
| 1.1984 | 0.4937 | 1000 | 1.2197 |
|
| 63 |
+
| 1.1556 | 0.5924 | 1200 | 1.2149 |
|
| 64 |
+
| 1.2386 | 0.6912 | 1400 | 1.2116 |
|
| 65 |
+
| 1.2661 | 0.7899 | 1600 | 1.2096 |
|
| 66 |
+
| 1.2752 | 0.8887 | 1800 | 1.2088 |
|
| 67 |
+
| 1.2701 | 0.9874 | 2000 | 1.2086 |
|
| 68 |
+
|
| 69 |
+
|
| 70 |
+
### Framework versions
|
| 71 |
+
|
| 72 |
+
- PEFT 0.7.1
|
| 73 |
+
- Transformers 4.40.1
|
| 74 |
+
- Pytorch 2.2.1+cu121
|
| 75 |
+
- Datasets 2.19.0
|
| 76 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 0.9997531473710195,
|
| 3 |
+
"total_flos": 4526278881050624.0,
|
| 4 |
+
"train_loss": 1.270192005722611,
|
| 5 |
+
"train_runtime": 113933.3906,
|
| 6 |
+
"train_samples": 48610,
|
| 7 |
+
"train_samples_per_second": 0.427,
|
| 8 |
+
"train_steps_per_second": 0.018
|
| 9 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 0.9997531473710195,
|
| 3 |
+
"total_flos": 4526278881050624.0,
|
| 4 |
+
"train_loss": 1.270192005722611,
|
| 5 |
+
"train_runtime": 113933.3906,
|
| 6 |
+
"train_samples": 48610,
|
| 7 |
+
"train_samples_per_second": 0.427,
|
| 8 |
+
"train_steps_per_second": 0.018
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2952 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 1.208633542060852,
|
| 3 |
+
"best_model_checkpoint": "/mnt/users/n3thakur/vectara/huggingface-dpo/trained_models/v3/Mistral-7B-Instruct-v0.2-miracl-raft-sft-v2.0/checkpoint-2000",
|
| 4 |
+
"epoch": 0.9997531473710195,
|
| 5 |
+
"eval_steps": 200,
|
| 6 |
+
"global_step": 2025,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0004937052579609973,
|
| 13 |
+
"grad_norm": 3.3225639243513934,
|
| 14 |
+
"learning_rate": 4.926108374384237e-08,
|
| 15 |
+
"loss": 1.6639,
|
| 16 |
+
"step": 1
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.0024685262898049864,
|
| 20 |
+
"grad_norm": 2.953462068009019,
|
| 21 |
+
"learning_rate": 2.4630541871921185e-07,
|
| 22 |
+
"loss": 1.7388,
|
| 23 |
+
"step": 5
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.004937052579609973,
|
| 27 |
+
"grad_norm": 2.4057426602070646,
|
| 28 |
+
"learning_rate": 4.926108374384237e-07,
|
| 29 |
+
"loss": 1.5798,
|
| 30 |
+
"step": 10
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.00740557886941496,
|
| 34 |
+
"grad_norm": 3.186500367639329,
|
| 35 |
+
"learning_rate": 7.389162561576356e-07,
|
| 36 |
+
"loss": 1.659,
|
| 37 |
+
"step": 15
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.009874105159219946,
|
| 41 |
+
"grad_norm": 2.396847476828635,
|
| 42 |
+
"learning_rate": 9.852216748768474e-07,
|
| 43 |
+
"loss": 1.6374,
|
| 44 |
+
"step": 20
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.012342631449024932,
|
| 48 |
+
"grad_norm": 2.5391731431636106,
|
| 49 |
+
"learning_rate": 1.2315270935960593e-06,
|
| 50 |
+
"loss": 1.6314,
|
| 51 |
+
"step": 25
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.01481115773882992,
|
| 55 |
+
"grad_norm": 2.2574409610476462,
|
| 56 |
+
"learning_rate": 1.4778325123152712e-06,
|
| 57 |
+
"loss": 1.5888,
|
| 58 |
+
"step": 30
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.017279684028634903,
|
| 62 |
+
"grad_norm": 1.8342813542038656,
|
| 63 |
+
"learning_rate": 1.724137931034483e-06,
|
| 64 |
+
"loss": 1.5412,
|
| 65 |
+
"step": 35
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.01974821031843989,
|
| 69 |
+
"grad_norm": 1.8380484598711315,
|
| 70 |
+
"learning_rate": 1.970443349753695e-06,
|
| 71 |
+
"loss": 1.4889,
|
| 72 |
+
"step": 40
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.02221673660824488,
|
| 76 |
+
"grad_norm": 1.613058249947001,
|
| 77 |
+
"learning_rate": 2.2167487684729067e-06,
|
| 78 |
+
"loss": 1.5403,
|
| 79 |
+
"step": 45
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.024685262898049863,
|
| 83 |
+
"grad_norm": 1.8920016704605567,
|
| 84 |
+
"learning_rate": 2.4630541871921186e-06,
|
| 85 |
+
"loss": 1.4831,
|
| 86 |
+
"step": 50
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.02715378918785485,
|
| 90 |
+
"grad_norm": 1.2203038504329438,
|
| 91 |
+
"learning_rate": 2.70935960591133e-06,
|
| 92 |
+
"loss": 1.4233,
|
| 93 |
+
"step": 55
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.02962231547765984,
|
| 97 |
+
"grad_norm": 1.095184752565883,
|
| 98 |
+
"learning_rate": 2.9556650246305424e-06,
|
| 99 |
+
"loss": 1.4255,
|
| 100 |
+
"step": 60
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.03209084176746482,
|
| 104 |
+
"grad_norm": 0.8448900941377993,
|
| 105 |
+
"learning_rate": 3.201970443349754e-06,
|
| 106 |
+
"loss": 1.4892,
|
| 107 |
+
"step": 65
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.03455936805726981,
|
| 111 |
+
"grad_norm": 0.9405655862570454,
|
| 112 |
+
"learning_rate": 3.448275862068966e-06,
|
| 113 |
+
"loss": 1.4673,
|
| 114 |
+
"step": 70
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.0370278943470748,
|
| 118 |
+
"grad_norm": 0.6713923929675227,
|
| 119 |
+
"learning_rate": 3.6945812807881777e-06,
|
| 120 |
+
"loss": 1.4148,
|
| 121 |
+
"step": 75
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.03949642063687978,
|
| 125 |
+
"grad_norm": 0.7755902373813679,
|
| 126 |
+
"learning_rate": 3.94088669950739e-06,
|
| 127 |
+
"loss": 1.4867,
|
| 128 |
+
"step": 80
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.04196494692668477,
|
| 132 |
+
"grad_norm": 0.7400218273582495,
|
| 133 |
+
"learning_rate": 4.1871921182266015e-06,
|
| 134 |
+
"loss": 1.3834,
|
| 135 |
+
"step": 85
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.04443347321648976,
|
| 139 |
+
"grad_norm": 0.7245551973919236,
|
| 140 |
+
"learning_rate": 4.4334975369458135e-06,
|
| 141 |
+
"loss": 1.44,
|
| 142 |
+
"step": 90
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.04690199950629474,
|
| 146 |
+
"grad_norm": 0.6731482962094358,
|
| 147 |
+
"learning_rate": 4.6798029556650245e-06,
|
| 148 |
+
"loss": 1.362,
|
| 149 |
+
"step": 95
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.049370525796099726,
|
| 153 |
+
"grad_norm": 0.7105341248736622,
|
| 154 |
+
"learning_rate": 4.926108374384237e-06,
|
| 155 |
+
"loss": 1.3716,
|
| 156 |
+
"step": 100
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.05183905208590472,
|
| 160 |
+
"grad_norm": 0.6774223469533757,
|
| 161 |
+
"learning_rate": 5.172413793103449e-06,
|
| 162 |
+
"loss": 1.4056,
|
| 163 |
+
"step": 105
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.0543075783757097,
|
| 167 |
+
"grad_norm": 0.6745745206164803,
|
| 168 |
+
"learning_rate": 5.41871921182266e-06,
|
| 169 |
+
"loss": 1.337,
|
| 170 |
+
"step": 110
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.056776104665514686,
|
| 174 |
+
"grad_norm": 0.5935854583319804,
|
| 175 |
+
"learning_rate": 5.665024630541872e-06,
|
| 176 |
+
"loss": 1.3615,
|
| 177 |
+
"step": 115
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.05924463095531968,
|
| 181 |
+
"grad_norm": 0.556924082351685,
|
| 182 |
+
"learning_rate": 5.911330049261085e-06,
|
| 183 |
+
"loss": 1.4248,
|
| 184 |
+
"step": 120
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.06171315724512466,
|
| 188 |
+
"grad_norm": 0.5427807889259738,
|
| 189 |
+
"learning_rate": 6.157635467980296e-06,
|
| 190 |
+
"loss": 1.3286,
|
| 191 |
+
"step": 125
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.06418168353492965,
|
| 195 |
+
"grad_norm": 0.6063519166723843,
|
| 196 |
+
"learning_rate": 6.403940886699508e-06,
|
| 197 |
+
"loss": 1.3176,
|
| 198 |
+
"step": 130
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.06665020982473463,
|
| 202 |
+
"grad_norm": 0.5670363529677274,
|
| 203 |
+
"learning_rate": 6.65024630541872e-06,
|
| 204 |
+
"loss": 1.3273,
|
| 205 |
+
"step": 135
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.06911873611453961,
|
| 209 |
+
"grad_norm": 0.5846835420330245,
|
| 210 |
+
"learning_rate": 6.896551724137932e-06,
|
| 211 |
+
"loss": 1.326,
|
| 212 |
+
"step": 140
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.07158726240434461,
|
| 216 |
+
"grad_norm": 0.5686293276495719,
|
| 217 |
+
"learning_rate": 7.1428571428571436e-06,
|
| 218 |
+
"loss": 1.376,
|
| 219 |
+
"step": 145
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.0740557886941496,
|
| 223 |
+
"grad_norm": 0.5275299029056365,
|
| 224 |
+
"learning_rate": 7.3891625615763555e-06,
|
| 225 |
+
"loss": 1.3364,
|
| 226 |
+
"step": 150
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.07652431498395458,
|
| 230 |
+
"grad_norm": 0.5464387846115857,
|
| 231 |
+
"learning_rate": 7.635467980295567e-06,
|
| 232 |
+
"loss": 1.3654,
|
| 233 |
+
"step": 155
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.07899284127375956,
|
| 237 |
+
"grad_norm": 0.5229203264129956,
|
| 238 |
+
"learning_rate": 7.88177339901478e-06,
|
| 239 |
+
"loss": 1.3009,
|
| 240 |
+
"step": 160
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.08146136756356455,
|
| 244 |
+
"grad_norm": 0.5746356087172889,
|
| 245 |
+
"learning_rate": 8.12807881773399e-06,
|
| 246 |
+
"loss": 1.2611,
|
| 247 |
+
"step": 165
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.08392989385336953,
|
| 251 |
+
"grad_norm": 0.5922232695792946,
|
| 252 |
+
"learning_rate": 8.374384236453203e-06,
|
| 253 |
+
"loss": 1.3643,
|
| 254 |
+
"step": 170
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.08639842014317453,
|
| 258 |
+
"grad_norm": 0.5295655281983137,
|
| 259 |
+
"learning_rate": 8.620689655172414e-06,
|
| 260 |
+
"loss": 1.3165,
|
| 261 |
+
"step": 175
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.08886694643297952,
|
| 265 |
+
"grad_norm": 0.5850731545805168,
|
| 266 |
+
"learning_rate": 8.866995073891627e-06,
|
| 267 |
+
"loss": 1.3105,
|
| 268 |
+
"step": 180
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.0913354727227845,
|
| 272 |
+
"grad_norm": 0.5551320012809824,
|
| 273 |
+
"learning_rate": 9.113300492610838e-06,
|
| 274 |
+
"loss": 1.278,
|
| 275 |
+
"step": 185
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.09380399901258948,
|
| 279 |
+
"grad_norm": 0.5711739398485313,
|
| 280 |
+
"learning_rate": 9.359605911330049e-06,
|
| 281 |
+
"loss": 1.3197,
|
| 282 |
+
"step": 190
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.09627252530239447,
|
| 286 |
+
"grad_norm": 0.5559894427552352,
|
| 287 |
+
"learning_rate": 9.605911330049262e-06,
|
| 288 |
+
"loss": 1.3409,
|
| 289 |
+
"step": 195
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.09874105159219945,
|
| 293 |
+
"grad_norm": 0.5671580477100892,
|
| 294 |
+
"learning_rate": 9.852216748768475e-06,
|
| 295 |
+
"loss": 1.3095,
|
| 296 |
+
"step": 200
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.09874105159219945,
|
| 300 |
+
"eval_loss": 1.2800045013427734,
|
| 301 |
+
"eval_runtime": 2727.7043,
|
| 302 |
+
"eval_samples_per_second": 1.466,
|
| 303 |
+
"eval_steps_per_second": 0.122,
|
| 304 |
+
"step": 200
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 0.10120957788200444,
|
| 308 |
+
"grad_norm": 0.5279833943359168,
|
| 309 |
+
"learning_rate": 9.999970269475589e-06,
|
| 310 |
+
"loss": 1.2966,
|
| 311 |
+
"step": 205
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 0.10367810417180943,
|
| 315 |
+
"grad_norm": 0.5644566149733632,
|
| 316 |
+
"learning_rate": 9.99963580513638e-06,
|
| 317 |
+
"loss": 1.2874,
|
| 318 |
+
"step": 210
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 0.10614663046161442,
|
| 322 |
+
"grad_norm": 0.6258961143244912,
|
| 323 |
+
"learning_rate": 9.998929738244678e-06,
|
| 324 |
+
"loss": 1.3209,
|
| 325 |
+
"step": 215
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.1086151567514194,
|
| 329 |
+
"grad_norm": 0.4834621448531187,
|
| 330 |
+
"learning_rate": 9.997852121279563e-06,
|
| 331 |
+
"loss": 1.3313,
|
| 332 |
+
"step": 220
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.11108368304122439,
|
| 336 |
+
"grad_norm": 0.5481752837030147,
|
| 337 |
+
"learning_rate": 9.996403034335912e-06,
|
| 338 |
+
"loss": 1.2738,
|
| 339 |
+
"step": 225
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.11355220933102937,
|
| 343 |
+
"grad_norm": 0.5886589355414898,
|
| 344 |
+
"learning_rate": 9.994582585118449e-06,
|
| 345 |
+
"loss": 1.2758,
|
| 346 |
+
"step": 230
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 0.11602073562083436,
|
| 350 |
+
"grad_norm": 0.5757139355018718,
|
| 351 |
+
"learning_rate": 9.992390908933746e-06,
|
| 352 |
+
"loss": 1.3187,
|
| 353 |
+
"step": 235
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 0.11848926191063935,
|
| 357 |
+
"grad_norm": 0.5464825333851621,
|
| 358 |
+
"learning_rate": 9.989828168680164e-06,
|
| 359 |
+
"loss": 1.3677,
|
| 360 |
+
"step": 240
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 0.12095778820044434,
|
| 364 |
+
"grad_norm": 0.6372982363888493,
|
| 365 |
+
"learning_rate": 9.986894554835735e-06,
|
| 366 |
+
"loss": 1.2668,
|
| 367 |
+
"step": 245
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 0.12342631449024932,
|
| 371 |
+
"grad_norm": 0.5445141174147589,
|
| 372 |
+
"learning_rate": 9.983590285444025e-06,
|
| 373 |
+
"loss": 1.2917,
|
| 374 |
+
"step": 250
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.1258948407800543,
|
| 378 |
+
"grad_norm": 0.6832031232821291,
|
| 379 |
+
"learning_rate": 9.979915606097907e-06,
|
| 380 |
+
"loss": 1.2675,
|
| 381 |
+
"step": 255
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 0.1283633670698593,
|
| 385 |
+
"grad_norm": 0.62128138673847,
|
| 386 |
+
"learning_rate": 9.975870789921322e-06,
|
| 387 |
+
"loss": 1.3187,
|
| 388 |
+
"step": 260
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 0.13083189335966428,
|
| 392 |
+
"grad_norm": 0.5161196413352727,
|
| 393 |
+
"learning_rate": 9.971456137548971e-06,
|
| 394 |
+
"loss": 1.3031,
|
| 395 |
+
"step": 265
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 0.13330041964946926,
|
| 399 |
+
"grad_norm": 0.5524745641605668,
|
| 400 |
+
"learning_rate": 9.966671977103972e-06,
|
| 401 |
+
"loss": 1.2749,
|
| 402 |
+
"step": 270
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 0.13576894593927424,
|
| 406 |
+
"grad_norm": 0.6669242272051678,
|
| 407 |
+
"learning_rate": 9.961518664173473e-06,
|
| 408 |
+
"loss": 1.3409,
|
| 409 |
+
"step": 275
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 0.13823747222907923,
|
| 413 |
+
"grad_norm": 0.5555562003933405,
|
| 414 |
+
"learning_rate": 9.955996581782218e-06,
|
| 415 |
+
"loss": 1.2468,
|
| 416 |
+
"step": 280
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.14070599851888424,
|
| 420 |
+
"grad_norm": 0.6244202172570701,
|
| 421 |
+
"learning_rate": 9.950106140364089e-06,
|
| 422 |
+
"loss": 1.3318,
|
| 423 |
+
"step": 285
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 0.14317452480868922,
|
| 427 |
+
"grad_norm": 0.5100271270558925,
|
| 428 |
+
"learning_rate": 9.943847777731584e-06,
|
| 429 |
+
"loss": 1.2522,
|
| 430 |
+
"step": 290
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 0.1456430510984942,
|
| 434 |
+
"grad_norm": 0.5482368116306139,
|
| 435 |
+
"learning_rate": 9.937221959043294e-06,
|
| 436 |
+
"loss": 1.3044,
|
| 437 |
+
"step": 295
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 0.1481115773882992,
|
| 441 |
+
"grad_norm": 0.5919271032213149,
|
| 442 |
+
"learning_rate": 9.93022917676932e-06,
|
| 443 |
+
"loss": 1.3131,
|
| 444 |
+
"step": 300
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 0.15058010367810418,
|
| 448 |
+
"grad_norm": 0.5428829828459178,
|
| 449 |
+
"learning_rate": 9.922869950654662e-06,
|
| 450 |
+
"loss": 1.2306,
|
| 451 |
+
"step": 305
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 0.15304862996790916,
|
| 455 |
+
"grad_norm": 0.5461192699131175,
|
| 456 |
+
"learning_rate": 9.915144827680606e-06,
|
| 457 |
+
"loss": 1.3151,
|
| 458 |
+
"step": 310
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.15551715625771415,
|
| 462 |
+
"grad_norm": 0.5113904915941117,
|
| 463 |
+
"learning_rate": 9.907054382024058e-06,
|
| 464 |
+
"loss": 1.2813,
|
| 465 |
+
"step": 315
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 0.15798568254751913,
|
| 469 |
+
"grad_norm": 0.6272053783824121,
|
| 470 |
+
"learning_rate": 9.898599215014868e-06,
|
| 471 |
+
"loss": 1.3064,
|
| 472 |
+
"step": 320
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 0.1604542088373241,
|
| 476 |
+
"grad_norm": 0.5671094073178861,
|
| 477 |
+
"learning_rate": 9.889779955091142e-06,
|
| 478 |
+
"loss": 1.2734,
|
| 479 |
+
"step": 325
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 0.1629227351271291,
|
| 483 |
+
"grad_norm": 0.582371136771928,
|
| 484 |
+
"learning_rate": 9.880597257752522e-06,
|
| 485 |
+
"loss": 1.3075,
|
| 486 |
+
"step": 330
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 0.16539126141693408,
|
| 490 |
+
"grad_norm": 0.5520015589132342,
|
| 491 |
+
"learning_rate": 9.87105180551148e-06,
|
| 492 |
+
"loss": 1.2802,
|
| 493 |
+
"step": 335
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 0.16785978770673907,
|
| 497 |
+
"grad_norm": 0.5937587353133906,
|
| 498 |
+
"learning_rate": 9.861144307842574e-06,
|
| 499 |
+
"loss": 1.2893,
|
| 500 |
+
"step": 340
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 0.17032831399654405,
|
| 504 |
+
"grad_norm": 0.5371728696508287,
|
| 505 |
+
"learning_rate": 9.850875501129726e-06,
|
| 506 |
+
"loss": 1.219,
|
| 507 |
+
"step": 345
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 0.17279684028634906,
|
| 511 |
+
"grad_norm": 0.5892603164875664,
|
| 512 |
+
"learning_rate": 9.840246148611485e-06,
|
| 513 |
+
"loss": 1.3094,
|
| 514 |
+
"step": 350
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 0.17526536657615405,
|
| 518 |
+
"grad_norm": 0.5502008403202052,
|
| 519 |
+
"learning_rate": 9.829257040324308e-06,
|
| 520 |
+
"loss": 1.2543,
|
| 521 |
+
"step": 355
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 0.17773389286595903,
|
| 525 |
+
"grad_norm": 0.6273336128612022,
|
| 526 |
+
"learning_rate": 9.817908993043819e-06,
|
| 527 |
+
"loss": 1.3107,
|
| 528 |
+
"step": 360
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 0.18020241915576402,
|
| 532 |
+
"grad_norm": 0.5761032807193177,
|
| 533 |
+
"learning_rate": 9.806202850224123e-06,
|
| 534 |
+
"loss": 1.2657,
|
| 535 |
+
"step": 365
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 0.182670945445569,
|
| 539 |
+
"grad_norm": 0.5628854954179761,
|
| 540 |
+
"learning_rate": 9.794139481935108e-06,
|
| 541 |
+
"loss": 1.258,
|
| 542 |
+
"step": 370
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 0.18513947173537398,
|
| 546 |
+
"grad_norm": 0.5637909618250402,
|
| 547 |
+
"learning_rate": 9.781719784797773e-06,
|
| 548 |
+
"loss": 1.2406,
|
| 549 |
+
"step": 375
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 0.18760799802517897,
|
| 553 |
+
"grad_norm": 0.5212794091813217,
|
| 554 |
+
"learning_rate": 9.768944681917582e-06,
|
| 555 |
+
"loss": 1.2391,
|
| 556 |
+
"step": 380
|
| 557 |
+
},
|
| 558 |
+
{
|
| 559 |
+
"epoch": 0.19007652431498395,
|
| 560 |
+
"grad_norm": 0.6416799620777229,
|
| 561 |
+
"learning_rate": 9.755815122815871e-06,
|
| 562 |
+
"loss": 1.3188,
|
| 563 |
+
"step": 385
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 0.19254505060478894,
|
| 567 |
+
"grad_norm": 0.5487444911675088,
|
| 568 |
+
"learning_rate": 9.742332083359252e-06,
|
| 569 |
+
"loss": 1.2884,
|
| 570 |
+
"step": 390
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 0.19501357689459392,
|
| 574 |
+
"grad_norm": 0.5697317991057302,
|
| 575 |
+
"learning_rate": 9.728496565687096e-06,
|
| 576 |
+
"loss": 1.2798,
|
| 577 |
+
"step": 395
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 0.1974821031843989,
|
| 581 |
+
"grad_norm": 0.6703007559314436,
|
| 582 |
+
"learning_rate": 9.714309598137045e-06,
|
| 583 |
+
"loss": 1.249,
|
| 584 |
+
"step": 400
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 0.1974821031843989,
|
| 588 |
+
"eval_loss": 1.2516121864318848,
|
| 589 |
+
"eval_runtime": 2575.7168,
|
| 590 |
+
"eval_samples_per_second": 1.553,
|
| 591 |
+
"eval_steps_per_second": 0.13,
|
| 592 |
+
"step": 400
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 0.1999506294742039,
|
| 596 |
+
"grad_norm": 0.526231295870319,
|
| 597 |
+
"learning_rate": 9.699772235168572e-06,
|
| 598 |
+
"loss": 1.2554,
|
| 599 |
+
"step": 405
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 0.20241915576400887,
|
| 603 |
+
"grad_norm": 0.5513334850915074,
|
| 604 |
+
"learning_rate": 9.68488555728462e-06,
|
| 605 |
+
"loss": 1.2753,
|
| 606 |
+
"step": 410
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 0.20488768205381389,
|
| 610 |
+
"grad_norm": 0.5979774809603526,
|
| 611 |
+
"learning_rate": 9.669650670951282e-06,
|
| 612 |
+
"loss": 1.2562,
|
| 613 |
+
"step": 415
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 0.20735620834361887,
|
| 617 |
+
"grad_norm": 0.5596269907913185,
|
| 618 |
+
"learning_rate": 9.654068708515564e-06,
|
| 619 |
+
"loss": 1.2829,
|
| 620 |
+
"step": 420
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 0.20982473463342385,
|
| 624 |
+
"grad_norm": 0.5593282633769885,
|
| 625 |
+
"learning_rate": 9.638140828121232e-06,
|
| 626 |
+
"loss": 1.2843,
|
| 627 |
+
"step": 425
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 0.21229326092322884,
|
| 631 |
+
"grad_norm": 0.5775937654131708,
|
| 632 |
+
"learning_rate": 9.621868213622713e-06,
|
| 633 |
+
"loss": 1.3001,
|
| 634 |
+
"step": 430
|
| 635 |
+
},
|
| 636 |
+
{
|
| 637 |
+
"epoch": 0.21476178721303382,
|
| 638 |
+
"grad_norm": 0.5661901033745343,
|
| 639 |
+
"learning_rate": 9.605252074497125e-06,
|
| 640 |
+
"loss": 1.3038,
|
| 641 |
+
"step": 435
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 0.2172303135028388,
|
| 645 |
+
"grad_norm": 0.6132749209816828,
|
| 646 |
+
"learning_rate": 9.588293645754363e-06,
|
| 647 |
+
"loss": 1.2843,
|
| 648 |
+
"step": 440
|
| 649 |
+
},
|
| 650 |
+
{
|
| 651 |
+
"epoch": 0.2196988397926438,
|
| 652 |
+
"grad_norm": 0.5624360623535388,
|
| 653 |
+
"learning_rate": 9.570994187845323e-06,
|
| 654 |
+
"loss": 1.2342,
|
| 655 |
+
"step": 445
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 0.22216736608244878,
|
| 659 |
+
"grad_norm": 0.5567610470805882,
|
| 660 |
+
"learning_rate": 9.553354986568201e-06,
|
| 661 |
+
"loss": 1.2955,
|
| 662 |
+
"step": 450
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 0.22463589237225376,
|
| 666 |
+
"grad_norm": 0.6255724221196046,
|
| 667 |
+
"learning_rate": 9.53537735297294e-06,
|
| 668 |
+
"loss": 1.2921,
|
| 669 |
+
"step": 455
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 0.22710441866205874,
|
| 673 |
+
"grad_norm": 0.5322242379012073,
|
| 674 |
+
"learning_rate": 9.517062623263768e-06,
|
| 675 |
+
"loss": 1.3011,
|
| 676 |
+
"step": 460
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"epoch": 0.22957294495186373,
|
| 680 |
+
"grad_norm": 0.5444205798338807,
|
| 681 |
+
"learning_rate": 9.498412158699905e-06,
|
| 682 |
+
"loss": 1.2733,
|
| 683 |
+
"step": 465
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"epoch": 0.2320414712416687,
|
| 687 |
+
"grad_norm": 0.5426713243893322,
|
| 688 |
+
"learning_rate": 9.479427345494366e-06,
|
| 689 |
+
"loss": 1.2312,
|
| 690 |
+
"step": 470
|
| 691 |
+
},
|
| 692 |
+
{
|
| 693 |
+
"epoch": 0.23450999753147372,
|
| 694 |
+
"grad_norm": 0.5871783813919782,
|
| 695 |
+
"learning_rate": 9.460109594710942e-06,
|
| 696 |
+
"loss": 1.3655,
|
| 697 |
+
"step": 475
|
| 698 |
+
},
|
| 699 |
+
{
|
| 700 |
+
"epoch": 0.2369785238212787,
|
| 701 |
+
"grad_norm": 0.574852380091512,
|
| 702 |
+
"learning_rate": 9.440460342159314e-06,
|
| 703 |
+
"loss": 1.2915,
|
| 704 |
+
"step": 480
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 0.2394470501110837,
|
| 708 |
+
"grad_norm": 0.5336092545421678,
|
| 709 |
+
"learning_rate": 9.42048104828834e-06,
|
| 710 |
+
"loss": 1.2963,
|
| 711 |
+
"step": 485
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 0.24191557640088868,
|
| 715 |
+
"grad_norm": 0.5998428802300876,
|
| 716 |
+
"learning_rate": 9.40017319807751e-06,
|
| 717 |
+
"loss": 1.3058,
|
| 718 |
+
"step": 490
|
| 719 |
+
},
|
| 720 |
+
{
|
| 721 |
+
"epoch": 0.24438410269069366,
|
| 722 |
+
"grad_norm": 0.5421507806800733,
|
| 723 |
+
"learning_rate": 9.379538300926553e-06,
|
| 724 |
+
"loss": 1.2881,
|
| 725 |
+
"step": 495
|
| 726 |
+
},
|
| 727 |
+
{
|
| 728 |
+
"epoch": 0.24685262898049865,
|
| 729 |
+
"grad_norm": 0.5358621498972941,
|
| 730 |
+
"learning_rate": 9.358577890543277e-06,
|
| 731 |
+
"loss": 1.2602,
|
| 732 |
+
"step": 500
|
| 733 |
+
},
|
| 734 |
+
{
|
| 735 |
+
"epoch": 0.24932115527030363,
|
| 736 |
+
"grad_norm": 0.564112204428148,
|
| 737 |
+
"learning_rate": 9.33729352482956e-06,
|
| 738 |
+
"loss": 1.279,
|
| 739 |
+
"step": 505
|
| 740 |
+
},
|
| 741 |
+
{
|
| 742 |
+
"epoch": 0.2517896815601086,
|
| 743 |
+
"grad_norm": 0.6382679375882034,
|
| 744 |
+
"learning_rate": 9.315686785765556e-06,
|
| 745 |
+
"loss": 1.2534,
|
| 746 |
+
"step": 510
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 0.2542582078499136,
|
| 750 |
+
"grad_norm": 0.5744585475791394,
|
| 751 |
+
"learning_rate": 9.293759279292116e-06,
|
| 752 |
+
"loss": 1.2744,
|
| 753 |
+
"step": 515
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 0.2567267341397186,
|
| 757 |
+
"grad_norm": 0.615942623926986,
|
| 758 |
+
"learning_rate": 9.271512635191427e-06,
|
| 759 |
+
"loss": 1.3055,
|
| 760 |
+
"step": 520
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"epoch": 0.25919526042952357,
|
| 764 |
+
"grad_norm": 0.5780670121734512,
|
| 765 |
+
"learning_rate": 9.248948506965877e-06,
|
| 766 |
+
"loss": 1.3175,
|
| 767 |
+
"step": 525
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 0.26166378671932855,
|
| 771 |
+
"grad_norm": 0.5777138377025286,
|
| 772 |
+
"learning_rate": 9.22606857171515e-06,
|
| 773 |
+
"loss": 1.2869,
|
| 774 |
+
"step": 530
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 0.26413231300913353,
|
| 778 |
+
"grad_norm": 0.5611724611846367,
|
| 779 |
+
"learning_rate": 9.202874530011583e-06,
|
| 780 |
+
"loss": 1.3199,
|
| 781 |
+
"step": 535
|
| 782 |
+
},
|
| 783 |
+
{
|
| 784 |
+
"epoch": 0.2666008392989385,
|
| 785 |
+
"grad_norm": 0.540794710590132,
|
| 786 |
+
"learning_rate": 9.179368105773768e-06,
|
| 787 |
+
"loss": 1.208,
|
| 788 |
+
"step": 540
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 0.2690693655887435,
|
| 792 |
+
"grad_norm": 0.5581497544995145,
|
| 793 |
+
"learning_rate": 9.155551046138408e-06,
|
| 794 |
+
"loss": 1.2638,
|
| 795 |
+
"step": 545
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 0.2715378918785485,
|
| 799 |
+
"grad_norm": 0.560865648598851,
|
| 800 |
+
"learning_rate": 9.131425121330477e-06,
|
| 801 |
+
"loss": 1.2629,
|
| 802 |
+
"step": 550
|
| 803 |
+
},
|
| 804 |
+
{
|
| 805 |
+
"epoch": 0.27400641816835347,
|
| 806 |
+
"grad_norm": 0.5458754463390333,
|
| 807 |
+
"learning_rate": 9.10699212453164e-06,
|
| 808 |
+
"loss": 1.2578,
|
| 809 |
+
"step": 555
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 0.27647494445815846,
|
| 813 |
+
"grad_norm": 0.5468153448281193,
|
| 814 |
+
"learning_rate": 9.082253871746962e-06,
|
| 815 |
+
"loss": 1.2488,
|
| 816 |
+
"step": 560
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"epoch": 0.27894347074796344,
|
| 820 |
+
"grad_norm": 0.6168084406611584,
|
| 821 |
+
"learning_rate": 9.057212201669952e-06,
|
| 822 |
+
"loss": 1.2931,
|
| 823 |
+
"step": 565
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 0.2814119970377685,
|
| 827 |
+
"grad_norm": 0.5767023372783159,
|
| 828 |
+
"learning_rate": 9.031868975545884e-06,
|
| 829 |
+
"loss": 1.2267,
|
| 830 |
+
"step": 570
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 0.28388052332757346,
|
| 834 |
+
"grad_norm": 0.5315895904457054,
|
| 835 |
+
"learning_rate": 9.006226077033464e-06,
|
| 836 |
+
"loss": 1.2463,
|
| 837 |
+
"step": 575
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 0.28634904961737845,
|
| 841 |
+
"grad_norm": 0.5616058952533509,
|
| 842 |
+
"learning_rate": 8.980285412064827e-06,
|
| 843 |
+
"loss": 1.287,
|
| 844 |
+
"step": 580
|
| 845 |
+
},
|
| 846 |
+
{
|
| 847 |
+
"epoch": 0.28881757590718343,
|
| 848 |
+
"grad_norm": 0.5746998443271042,
|
| 849 |
+
"learning_rate": 8.954048908703873e-06,
|
| 850 |
+
"loss": 1.2929,
|
| 851 |
+
"step": 585
|
| 852 |
+
},
|
| 853 |
+
{
|
| 854 |
+
"epoch": 0.2912861021969884,
|
| 855 |
+
"grad_norm": 0.5551746835964705,
|
| 856 |
+
"learning_rate": 8.92751851700297e-06,
|
| 857 |
+
"loss": 1.298,
|
| 858 |
+
"step": 590
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 0.2937546284867934,
|
| 862 |
+
"grad_norm": 0.578564867995815,
|
| 863 |
+
"learning_rate": 8.900696208857996e-06,
|
| 864 |
+
"loss": 1.2973,
|
| 865 |
+
"step": 595
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 0.2962231547765984,
|
| 869 |
+
"grad_norm": 0.5925663520696334,
|
| 870 |
+
"learning_rate": 8.873583977861802e-06,
|
| 871 |
+
"loss": 1.2514,
|
| 872 |
+
"step": 600
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 0.2962231547765984,
|
| 876 |
+
"eval_loss": 1.2368682622909546,
|
| 877 |
+
"eval_runtime": 2566.7596,
|
| 878 |
+
"eval_samples_per_second": 1.558,
|
| 879 |
+
"eval_steps_per_second": 0.13,
|
| 880 |
+
"step": 600
|
| 881 |
+
},
|
| 882 |
+
{
|
| 883 |
+
"epoch": 0.29869168106640337,
|
| 884 |
+
"grad_norm": 0.5605310856508363,
|
| 885 |
+
"learning_rate": 8.846183839156015e-06,
|
| 886 |
+
"loss": 1.286,
|
| 887 |
+
"step": 605
|
| 888 |
+
},
|
| 889 |
+
{
|
| 890 |
+
"epoch": 0.30116020735620835,
|
| 891 |
+
"grad_norm": 0.6632798685747615,
|
| 892 |
+
"learning_rate": 8.818497829281272e-06,
|
| 893 |
+
"loss": 1.2916,
|
| 894 |
+
"step": 610
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"epoch": 0.30362873364601334,
|
| 898 |
+
"grad_norm": 0.6145012170463651,
|
| 899 |
+
"learning_rate": 8.790528006025848e-06,
|
| 900 |
+
"loss": 1.2788,
|
| 901 |
+
"step": 615
|
| 902 |
+
},
|
| 903 |
+
{
|
| 904 |
+
"epoch": 0.3060972599358183,
|
| 905 |
+
"grad_norm": 0.6017170291600934,
|
| 906 |
+
"learning_rate": 8.762276448272709e-06,
|
| 907 |
+
"loss": 1.3156,
|
| 908 |
+
"step": 620
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 0.3085657862256233,
|
| 912 |
+
"grad_norm": 0.5728547538871892,
|
| 913 |
+
"learning_rate": 8.733745255844996e-06,
|
| 914 |
+
"loss": 1.2592,
|
| 915 |
+
"step": 625
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 0.3110343125154283,
|
| 919 |
+
"grad_norm": 0.558142508046803,
|
| 920 |
+
"learning_rate": 8.70493654934996e-06,
|
| 921 |
+
"loss": 1.309,
|
| 922 |
+
"step": 630
|
| 923 |
+
},
|
| 924 |
+
{
|
| 925 |
+
"epoch": 0.3135028388052333,
|
| 926 |
+
"grad_norm": 0.5596812007471911,
|
| 927 |
+
"learning_rate": 8.675852470021344e-06,
|
| 928 |
+
"loss": 1.2746,
|
| 929 |
+
"step": 635
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 0.31597136509503826,
|
| 933 |
+
"grad_norm": 0.5909265132847957,
|
| 934 |
+
"learning_rate": 8.646495179560221e-06,
|
| 935 |
+
"loss": 1.2686,
|
| 936 |
+
"step": 640
|
| 937 |
+
},
|
| 938 |
+
{
|
| 939 |
+
"epoch": 0.31843989138484324,
|
| 940 |
+
"grad_norm": 0.6185942591784858,
|
| 941 |
+
"learning_rate": 8.616866859974344e-06,
|
| 942 |
+
"loss": 1.2759,
|
| 943 |
+
"step": 645
|
| 944 |
+
},
|
| 945 |
+
{
|
| 946 |
+
"epoch": 0.3209084176746482,
|
| 947 |
+
"grad_norm": 0.6157204431679958,
|
| 948 |
+
"learning_rate": 8.586969713415949e-06,
|
| 949 |
+
"loss": 1.2957,
|
| 950 |
+
"step": 650
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 0.3233769439644532,
|
| 954 |
+
"grad_norm": 0.5974197754755597,
|
| 955 |
+
"learning_rate": 8.556805962018091e-06,
|
| 956 |
+
"loss": 1.27,
|
| 957 |
+
"step": 655
|
| 958 |
+
},
|
| 959 |
+
{
|
| 960 |
+
"epoch": 0.3258454702542582,
|
| 961 |
+
"grad_norm": 0.5389440161380957,
|
| 962 |
+
"learning_rate": 8.526377847729475e-06,
|
| 963 |
+
"loss": 1.2925,
|
| 964 |
+
"step": 660
|
| 965 |
+
},
|
| 966 |
+
{
|
| 967 |
+
"epoch": 0.3283139965440632,
|
| 968 |
+
"grad_norm": 0.5370983741740369,
|
| 969 |
+
"learning_rate": 8.495687632147817e-06,
|
| 970 |
+
"loss": 1.2522,
|
| 971 |
+
"step": 665
|
| 972 |
+
},
|
| 973 |
+
{
|
| 974 |
+
"epoch": 0.33078252283386816,
|
| 975 |
+
"grad_norm": 0.5639132359450145,
|
| 976 |
+
"learning_rate": 8.46473759635176e-06,
|
| 977 |
+
"loss": 1.2595,
|
| 978 |
+
"step": 670
|
| 979 |
+
},
|
| 980 |
+
{
|
| 981 |
+
"epoch": 0.33325104912367315,
|
| 982 |
+
"grad_norm": 0.5598705018251675,
|
| 983 |
+
"learning_rate": 8.433530040731321e-06,
|
| 984 |
+
"loss": 1.2746,
|
| 985 |
+
"step": 675
|
| 986 |
+
},
|
| 987 |
+
{
|
| 988 |
+
"epoch": 0.33571957541347813,
|
| 989 |
+
"grad_norm": 0.6303186487688077,
|
| 990 |
+
"learning_rate": 8.402067284816919e-06,
|
| 991 |
+
"loss": 1.2701,
|
| 992 |
+
"step": 680
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 0.3381881017032831,
|
| 996 |
+
"grad_norm": 0.562747309348665,
|
| 997 |
+
"learning_rate": 8.370351667106969e-06,
|
| 998 |
+
"loss": 1.2305,
|
| 999 |
+
"step": 685
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 0.3406566279930881,
|
| 1003 |
+
"grad_norm": 0.5720387765798051,
|
| 1004 |
+
"learning_rate": 8.338385544894073e-06,
|
| 1005 |
+
"loss": 1.2047,
|
| 1006 |
+
"step": 690
|
| 1007 |
+
},
|
| 1008 |
+
{
|
| 1009 |
+
"epoch": 0.3431251542828931,
|
| 1010 |
+
"grad_norm": 0.5465830505695308,
|
| 1011 |
+
"learning_rate": 8.306171294089808e-06,
|
| 1012 |
+
"loss": 1.2507,
|
| 1013 |
+
"step": 695
|
| 1014 |
+
},
|
| 1015 |
+
{
|
| 1016 |
+
"epoch": 0.3455936805726981,
|
| 1017 |
+
"grad_norm": 0.5572297207326813,
|
| 1018 |
+
"learning_rate": 8.273711309048145e-06,
|
| 1019 |
+
"loss": 1.2599,
|
| 1020 |
+
"step": 700
|
| 1021 |
+
},
|
| 1022 |
+
{
|
| 1023 |
+
"epoch": 0.3480622068625031,
|
| 1024 |
+
"grad_norm": 0.5916945311296786,
|
| 1025 |
+
"learning_rate": 8.241008002387474e-06,
|
| 1026 |
+
"loss": 1.2615,
|
| 1027 |
+
"step": 705
|
| 1028 |
+
},
|
| 1029 |
+
{
|
| 1030 |
+
"epoch": 0.3505307331523081,
|
| 1031 |
+
"grad_norm": 0.6326075200444886,
|
| 1032 |
+
"learning_rate": 8.208063804811293e-06,
|
| 1033 |
+
"loss": 1.2559,
|
| 1034 |
+
"step": 710
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 0.3529992594421131,
|
| 1038 |
+
"grad_norm": 0.6229843020575793,
|
| 1039 |
+
"learning_rate": 8.174881164927535e-06,
|
| 1040 |
+
"loss": 1.2652,
|
| 1041 |
+
"step": 715
|
| 1042 |
+
},
|
| 1043 |
+
{
|
| 1044 |
+
"epoch": 0.35546778573191806,
|
| 1045 |
+
"grad_norm": 0.5926153932237264,
|
| 1046 |
+
"learning_rate": 8.141462549066581e-06,
|
| 1047 |
+
"loss": 1.2423,
|
| 1048 |
+
"step": 720
|
| 1049 |
+
},
|
| 1050 |
+
{
|
| 1051 |
+
"epoch": 0.35793631202172305,
|
| 1052 |
+
"grad_norm": 0.5293071287095781,
|
| 1053 |
+
"learning_rate": 8.107810441097948e-06,
|
| 1054 |
+
"loss": 1.2185,
|
| 1055 |
+
"step": 725
|
| 1056 |
+
},
|
| 1057 |
+
{
|
| 1058 |
+
"epoch": 0.36040483831152803,
|
| 1059 |
+
"grad_norm": 0.5950082298726722,
|
| 1060 |
+
"learning_rate": 8.073927342245663e-06,
|
| 1061 |
+
"loss": 1.2458,
|
| 1062 |
+
"step": 730
|
| 1063 |
+
},
|
| 1064 |
+
{
|
| 1065 |
+
"epoch": 0.362873364601333,
|
| 1066 |
+
"grad_norm": 0.5437872955630408,
|
| 1067 |
+
"learning_rate": 8.039815770902368e-06,
|
| 1068 |
+
"loss": 1.2699,
|
| 1069 |
+
"step": 735
|
| 1070 |
+
},
|
| 1071 |
+
{
|
| 1072 |
+
"epoch": 0.365341890891138,
|
| 1073 |
+
"grad_norm": 0.5842632003875607,
|
| 1074 |
+
"learning_rate": 8.005478262442132e-06,
|
| 1075 |
+
"loss": 1.2489,
|
| 1076 |
+
"step": 740
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 0.367810417180943,
|
| 1080 |
+
"grad_norm": 0.5957543279120926,
|
| 1081 |
+
"learning_rate": 7.970917369032011e-06,
|
| 1082 |
+
"loss": 1.2808,
|
| 1083 |
+
"step": 745
|
| 1084 |
+
},
|
| 1085 |
+
{
|
| 1086 |
+
"epoch": 0.37027894347074797,
|
| 1087 |
+
"grad_norm": 0.5573632520708609,
|
| 1088 |
+
"learning_rate": 7.936135659442355e-06,
|
| 1089 |
+
"loss": 1.2394,
|
| 1090 |
+
"step": 750
|
| 1091 |
+
},
|
| 1092 |
+
{
|
| 1093 |
+
"epoch": 0.37274746976055295,
|
| 1094 |
+
"grad_norm": 0.5383442104756702,
|
| 1095 |
+
"learning_rate": 7.901135718855877e-06,
|
| 1096 |
+
"loss": 1.2584,
|
| 1097 |
+
"step": 755
|
| 1098 |
+
},
|
| 1099 |
+
{
|
| 1100 |
+
"epoch": 0.37521599605035794,
|
| 1101 |
+
"grad_norm": 0.5269547291918393,
|
| 1102 |
+
"learning_rate": 7.86592014867551e-06,
|
| 1103 |
+
"loss": 1.32,
|
| 1104 |
+
"step": 760
|
| 1105 |
+
},
|
| 1106 |
+
{
|
| 1107 |
+
"epoch": 0.3776845223401629,
|
| 1108 |
+
"grad_norm": 0.6059173481615415,
|
| 1109 |
+
"learning_rate": 7.830491566331063e-06,
|
| 1110 |
+
"loss": 1.2705,
|
| 1111 |
+
"step": 765
|
| 1112 |
+
},
|
| 1113 |
+
{
|
| 1114 |
+
"epoch": 0.3801530486299679,
|
| 1115 |
+
"grad_norm": 0.5905241537228486,
|
| 1116 |
+
"learning_rate": 7.794852605084661e-06,
|
| 1117 |
+
"loss": 1.2661,
|
| 1118 |
+
"step": 770
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 0.3826215749197729,
|
| 1122 |
+
"grad_norm": 0.6119492506708828,
|
| 1123 |
+
"learning_rate": 7.759005913835048e-06,
|
| 1124 |
+
"loss": 1.2573,
|
| 1125 |
+
"step": 775
|
| 1126 |
+
},
|
| 1127 |
+
{
|
| 1128 |
+
"epoch": 0.3850901012095779,
|
| 1129 |
+
"grad_norm": 0.6449864393640712,
|
| 1130 |
+
"learning_rate": 7.722954156920675e-06,
|
| 1131 |
+
"loss": 1.2681,
|
| 1132 |
+
"step": 780
|
| 1133 |
+
},
|
| 1134 |
+
{
|
| 1135 |
+
"epoch": 0.38755862749938286,
|
| 1136 |
+
"grad_norm": 0.5777516112864801,
|
| 1137 |
+
"learning_rate": 7.686700013921704e-06,
|
| 1138 |
+
"loss": 1.2999,
|
| 1139 |
+
"step": 785
|
| 1140 |
+
},
|
| 1141 |
+
{
|
| 1142 |
+
"epoch": 0.39002715378918784,
|
| 1143 |
+
"grad_norm": 0.5818063096150684,
|
| 1144 |
+
"learning_rate": 7.650246179460826e-06,
|
| 1145 |
+
"loss": 1.2842,
|
| 1146 |
+
"step": 790
|
| 1147 |
+
},
|
| 1148 |
+
{
|
| 1149 |
+
"epoch": 0.3924956800789928,
|
| 1150 |
+
"grad_norm": 0.5844315528318011,
|
| 1151 |
+
"learning_rate": 7.613595363002977e-06,
|
| 1152 |
+
"loss": 1.2995,
|
| 1153 |
+
"step": 795
|
| 1154 |
+
},
|
| 1155 |
+
{
|
| 1156 |
+
"epoch": 0.3949642063687978,
|
| 1157 |
+
"grad_norm": 0.5560255613889942,
|
| 1158 |
+
"learning_rate": 7.57675028865397e-06,
|
| 1159 |
+
"loss": 1.275,
|
| 1160 |
+
"step": 800
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 0.3949642063687978,
|
| 1164 |
+
"eval_loss": 1.2263342142105103,
|
| 1165 |
+
"eval_runtime": 2463.6634,
|
| 1166 |
+
"eval_samples_per_second": 1.624,
|
| 1167 |
+
"eval_steps_per_second": 0.136,
|
| 1168 |
+
"step": 800
|
| 1169 |
+
},
|
| 1170 |
+
{
|
| 1171 |
+
"epoch": 0.3974327326586028,
|
| 1172 |
+
"grad_norm": 0.5523940138743026,
|
| 1173 |
+
"learning_rate": 7.539713694958013e-06,
|
| 1174 |
+
"loss": 1.2202,
|
| 1175 |
+
"step": 805
|
| 1176 |
+
},
|
| 1177 |
+
{
|
| 1178 |
+
"epoch": 0.3999012589484078,
|
| 1179 |
+
"grad_norm": 0.5936001183365429,
|
| 1180 |
+
"learning_rate": 7.502488334694167e-06,
|
| 1181 |
+
"loss": 1.2444,
|
| 1182 |
+
"step": 810
|
| 1183 |
+
},
|
| 1184 |
+
{
|
| 1185 |
+
"epoch": 0.40236978523821276,
|
| 1186 |
+
"grad_norm": 0.6143038376732798,
|
| 1187 |
+
"learning_rate": 7.465076974671739e-06,
|
| 1188 |
+
"loss": 1.2032,
|
| 1189 |
+
"step": 815
|
| 1190 |
+
},
|
| 1191 |
+
{
|
| 1192 |
+
"epoch": 0.40483831152801775,
|
| 1193 |
+
"grad_norm": 0.5865451493919344,
|
| 1194 |
+
"learning_rate": 7.427482395524646e-06,
|
| 1195 |
+
"loss": 1.2733,
|
| 1196 |
+
"step": 820
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 0.4073068378178228,
|
| 1200 |
+
"grad_norm": 0.5980943581114722,
|
| 1201 |
+
"learning_rate": 7.389707391504728e-06,
|
| 1202 |
+
"loss": 1.2732,
|
| 1203 |
+
"step": 825
|
| 1204 |
+
},
|
| 1205 |
+
{
|
| 1206 |
+
"epoch": 0.40977536410762777,
|
| 1207 |
+
"grad_norm": 0.6323487686008166,
|
| 1208 |
+
"learning_rate": 7.35175477027408e-06,
|
| 1209 |
+
"loss": 1.244,
|
| 1210 |
+
"step": 830
|
| 1211 |
+
},
|
| 1212 |
+
{
|
| 1213 |
+
"epoch": 0.41224389039743276,
|
| 1214 |
+
"grad_norm": 0.6562081554973773,
|
| 1215 |
+
"learning_rate": 7.313627352696353e-06,
|
| 1216 |
+
"loss": 1.2642,
|
| 1217 |
+
"step": 835
|
| 1218 |
+
},
|
| 1219 |
+
{
|
| 1220 |
+
"epoch": 0.41471241668723774,
|
| 1221 |
+
"grad_norm": 0.5554470118072983,
|
| 1222 |
+
"learning_rate": 7.2753279726271e-06,
|
| 1223 |
+
"loss": 1.2556,
|
| 1224 |
+
"step": 840
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"epoch": 0.4171809429770427,
|
| 1228 |
+
"grad_norm": 0.5740654163988275,
|
| 1229 |
+
"learning_rate": 7.236859476703148e-06,
|
| 1230 |
+
"loss": 1.2292,
|
| 1231 |
+
"step": 845
|
| 1232 |
+
},
|
| 1233 |
+
{
|
| 1234 |
+
"epoch": 0.4196494692668477,
|
| 1235 |
+
"grad_norm": 0.6062582969566837,
|
| 1236 |
+
"learning_rate": 7.198224724131012e-06,
|
| 1237 |
+
"loss": 1.235,
|
| 1238 |
+
"step": 850
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 0.4221179955566527,
|
| 1242 |
+
"grad_norm": 0.5434614048201878,
|
| 1243 |
+
"learning_rate": 7.159426586474388e-06,
|
| 1244 |
+
"loss": 1.2224,
|
| 1245 |
+
"step": 855
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 0.4245865218464577,
|
| 1249 |
+
"grad_norm": 0.5254561702235886,
|
| 1250 |
+
"learning_rate": 7.120467947440719e-06,
|
| 1251 |
+
"loss": 1.2557,
|
| 1252 |
+
"step": 860
|
| 1253 |
+
},
|
| 1254 |
+
{
|
| 1255 |
+
"epoch": 0.42705504813626266,
|
| 1256 |
+
"grad_norm": 0.5713031391494172,
|
| 1257 |
+
"learning_rate": 7.081351702666863e-06,
|
| 1258 |
+
"loss": 1.2063,
|
| 1259 |
+
"step": 865
|
| 1260 |
+
},
|
| 1261 |
+
{
|
| 1262 |
+
"epoch": 0.42952357442606764,
|
| 1263 |
+
"grad_norm": 0.5969980245366532,
|
| 1264 |
+
"learning_rate": 7.042080759503866e-06,
|
| 1265 |
+
"loss": 1.2418,
|
| 1266 |
+
"step": 870
|
| 1267 |
+
},
|
| 1268 |
+
{
|
| 1269 |
+
"epoch": 0.43199210071587263,
|
| 1270 |
+
"grad_norm": 0.5718940130718101,
|
| 1271 |
+
"learning_rate": 7.00265803680088e-06,
|
| 1272 |
+
"loss": 1.2108,
|
| 1273 |
+
"step": 875
|
| 1274 |
+
},
|
| 1275 |
+
{
|
| 1276 |
+
"epoch": 0.4344606270056776,
|
| 1277 |
+
"grad_norm": 0.6045555591926912,
|
| 1278 |
+
"learning_rate": 6.963086464688209e-06,
|
| 1279 |
+
"loss": 1.2597,
|
| 1280 |
+
"step": 880
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 0.4369291532954826,
|
| 1284 |
+
"grad_norm": 0.5566709780037437,
|
| 1285 |
+
"learning_rate": 6.923368984359526e-06,
|
| 1286 |
+
"loss": 1.2174,
|
| 1287 |
+
"step": 885
|
| 1288 |
+
},
|
| 1289 |
+
{
|
| 1290 |
+
"epoch": 0.4393976795852876,
|
| 1291 |
+
"grad_norm": 0.5630200258106689,
|
| 1292 |
+
"learning_rate": 6.883508547853268e-06,
|
| 1293 |
+
"loss": 1.2244,
|
| 1294 |
+
"step": 890
|
| 1295 |
+
},
|
| 1296 |
+
{
|
| 1297 |
+
"epoch": 0.44186620587509257,
|
| 1298 |
+
"grad_norm": 0.5348314552481888,
|
| 1299 |
+
"learning_rate": 6.843508117833224e-06,
|
| 1300 |
+
"loss": 1.2687,
|
| 1301 |
+
"step": 895
|
| 1302 |
+
},
|
| 1303 |
+
{
|
| 1304 |
+
"epoch": 0.44433473216489755,
|
| 1305 |
+
"grad_norm": 0.49625311943608336,
|
| 1306 |
+
"learning_rate": 6.8033706673683276e-06,
|
| 1307 |
+
"loss": 1.1986,
|
| 1308 |
+
"step": 900
|
| 1309 |
+
},
|
| 1310 |
+
{
|
| 1311 |
+
"epoch": 0.44680325845470253,
|
| 1312 |
+
"grad_norm": 0.5542218838145379,
|
| 1313 |
+
"learning_rate": 6.763099179711685e-06,
|
| 1314 |
+
"loss": 1.2286,
|
| 1315 |
+
"step": 905
|
| 1316 |
+
},
|
| 1317 |
+
{
|
| 1318 |
+
"epoch": 0.4492717847445075,
|
| 1319 |
+
"grad_norm": 0.594098893943127,
|
| 1320 |
+
"learning_rate": 6.722696648078838e-06,
|
| 1321 |
+
"loss": 1.2335,
|
| 1322 |
+
"step": 910
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 0.4517403110343125,
|
| 1326 |
+
"grad_norm": 0.5478077068384012,
|
| 1327 |
+
"learning_rate": 6.682166075425298e-06,
|
| 1328 |
+
"loss": 1.264,
|
| 1329 |
+
"step": 915
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 0.4542088373241175,
|
| 1333 |
+
"grad_norm": 0.5727528301850252,
|
| 1334 |
+
"learning_rate": 6.641510474223338e-06,
|
| 1335 |
+
"loss": 1.226,
|
| 1336 |
+
"step": 920
|
| 1337 |
+
},
|
| 1338 |
+
{
|
| 1339 |
+
"epoch": 0.45667736361392247,
|
| 1340 |
+
"grad_norm": 0.5888269073825134,
|
| 1341 |
+
"learning_rate": 6.600732866238097e-06,
|
| 1342 |
+
"loss": 1.212,
|
| 1343 |
+
"step": 925
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 0.45914588990372746,
|
| 1347 |
+
"grad_norm": 0.5736288265128395,
|
| 1348 |
+
"learning_rate": 6.559836282302984e-06,
|
| 1349 |
+
"loss": 1.25,
|
| 1350 |
+
"step": 930
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 0.46161441619353244,
|
| 1354 |
+
"grad_norm": 0.6651036803926929,
|
| 1355 |
+
"learning_rate": 6.5188237620943965e-06,
|
| 1356 |
+
"loss": 1.2672,
|
| 1357 |
+
"step": 935
|
| 1358 |
+
},
|
| 1359 |
+
{
|
| 1360 |
+
"epoch": 0.4640829424833374,
|
| 1361 |
+
"grad_norm": 0.5547382454730273,
|
| 1362 |
+
"learning_rate": 6.477698353905808e-06,
|
| 1363 |
+
"loss": 1.2887,
|
| 1364 |
+
"step": 940
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 0.4665514687731424,
|
| 1368 |
+
"grad_norm": 0.5627833712727636,
|
| 1369 |
+
"learning_rate": 6.436463114421199e-06,
|
| 1370 |
+
"loss": 1.2674,
|
| 1371 |
+
"step": 945
|
| 1372 |
+
},
|
| 1373 |
+
{
|
| 1374 |
+
"epoch": 0.46901999506294745,
|
| 1375 |
+
"grad_norm": 0.5562108977867529,
|
| 1376 |
+
"learning_rate": 6.395121108487855e-06,
|
| 1377 |
+
"loss": 1.2973,
|
| 1378 |
+
"step": 950
|
| 1379 |
+
},
|
| 1380 |
+
{
|
| 1381 |
+
"epoch": 0.47148852135275243,
|
| 1382 |
+
"grad_norm": 0.5940300188918287,
|
| 1383 |
+
"learning_rate": 6.353675408888582e-06,
|
| 1384 |
+
"loss": 1.278,
|
| 1385 |
+
"step": 955
|
| 1386 |
+
},
|
| 1387 |
+
{
|
| 1388 |
+
"epoch": 0.4739570476425574,
|
| 1389 |
+
"grad_norm": 0.6499724681591359,
|
| 1390 |
+
"learning_rate": 6.312129096113313e-06,
|
| 1391 |
+
"loss": 1.242,
|
| 1392 |
+
"step": 960
|
| 1393 |
+
},
|
| 1394 |
+
{
|
| 1395 |
+
"epoch": 0.4764255739323624,
|
| 1396 |
+
"grad_norm": 0.5794092582819724,
|
| 1397 |
+
"learning_rate": 6.270485258130146e-06,
|
| 1398 |
+
"loss": 1.2263,
|
| 1399 |
+
"step": 965
|
| 1400 |
+
},
|
| 1401 |
+
{
|
| 1402 |
+
"epoch": 0.4788941002221674,
|
| 1403 |
+
"grad_norm": 0.5810005883829364,
|
| 1404 |
+
"learning_rate": 6.228746990155831e-06,
|
| 1405 |
+
"loss": 1.2166,
|
| 1406 |
+
"step": 970
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 0.48136262651197237,
|
| 1410 |
+
"grad_norm": 0.5523321758038612,
|
| 1411 |
+
"learning_rate": 6.186917394425715e-06,
|
| 1412 |
+
"loss": 1.2666,
|
| 1413 |
+
"step": 975
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"epoch": 0.48383115280177735,
|
| 1417 |
+
"grad_norm": 0.5353766340095819,
|
| 1418 |
+
"learning_rate": 6.144999579963164e-06,
|
| 1419 |
+
"loss": 1.2332,
|
| 1420 |
+
"step": 980
|
| 1421 |
+
},
|
| 1422 |
+
{
|
| 1423 |
+
"epoch": 0.48629967909158234,
|
| 1424 |
+
"grad_norm": 0.5962559333577797,
|
| 1425 |
+
"learning_rate": 6.102996662348485e-06,
|
| 1426 |
+
"loss": 1.2985,
|
| 1427 |
+
"step": 985
|
| 1428 |
+
},
|
| 1429 |
+
{
|
| 1430 |
+
"epoch": 0.4887682053813873,
|
| 1431 |
+
"grad_norm": 0.573508927377536,
|
| 1432 |
+
"learning_rate": 6.060911763487353e-06,
|
| 1433 |
+
"loss": 1.2353,
|
| 1434 |
+
"step": 990
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 0.4912367316711923,
|
| 1438 |
+
"grad_norm": 0.6190411186907346,
|
| 1439 |
+
"learning_rate": 6.0187480113787765e-06,
|
| 1440 |
+
"loss": 1.2668,
|
| 1441 |
+
"step": 995
|
| 1442 |
+
},
|
| 1443 |
+
{
|
| 1444 |
+
"epoch": 0.4937052579609973,
|
| 1445 |
+
"grad_norm": 0.537107101144104,
|
| 1446 |
+
"learning_rate": 5.976508539882604e-06,
|
| 1447 |
+
"loss": 1.1984,
|
| 1448 |
+
"step": 1000
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 0.4937052579609973,
|
| 1452 |
+
"eval_loss": 1.2196881771087646,
|
| 1453 |
+
"eval_runtime": 2373.8686,
|
| 1454 |
+
"eval_samples_per_second": 1.685,
|
| 1455 |
+
"eval_steps_per_second": 0.141,
|
| 1456 |
+
"step": 1000
|
| 1457 |
+
},
|
| 1458 |
+
{
|
| 1459 |
+
"epoch": 0.4961737842508023,
|
| 1460 |
+
"grad_norm": 0.5673334311067016,
|
| 1461 |
+
"learning_rate": 5.934196488486594e-06,
|
| 1462 |
+
"loss": 1.2573,
|
| 1463 |
+
"step": 1005
|
| 1464 |
+
},
|
| 1465 |
+
{
|
| 1466 |
+
"epoch": 0.49864231054060726,
|
| 1467 |
+
"grad_norm": 0.6141102747872601,
|
| 1468 |
+
"learning_rate": 5.891815002073081e-06,
|
| 1469 |
+
"loss": 1.2776,
|
| 1470 |
+
"step": 1010
|
| 1471 |
+
},
|
| 1472 |
+
{
|
| 1473 |
+
"epoch": 0.5011108368304122,
|
| 1474 |
+
"grad_norm": 0.5866475421501153,
|
| 1475 |
+
"learning_rate": 5.849367230685214e-06,
|
| 1476 |
+
"loss": 1.2139,
|
| 1477 |
+
"step": 1015
|
| 1478 |
+
},
|
| 1479 |
+
{
|
| 1480 |
+
"epoch": 0.5035793631202172,
|
| 1481 |
+
"grad_norm": 0.5973223110810923,
|
| 1482 |
+
"learning_rate": 5.806856329292839e-06,
|
| 1483 |
+
"loss": 1.2809,
|
| 1484 |
+
"step": 1020
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 0.5060478894100222,
|
| 1488 |
+
"grad_norm": 0.6385978269750231,
|
| 1489 |
+
"learning_rate": 5.764285457557994e-06,
|
| 1490 |
+
"loss": 1.2511,
|
| 1491 |
+
"step": 1025
|
| 1492 |
+
},
|
| 1493 |
+
{
|
| 1494 |
+
"epoch": 0.5085164156998272,
|
| 1495 |
+
"grad_norm": 0.5607340345191899,
|
| 1496 |
+
"learning_rate": 5.721657779600071e-06,
|
| 1497 |
+
"loss": 1.2421,
|
| 1498 |
+
"step": 1030
|
| 1499 |
+
},
|
| 1500 |
+
{
|
| 1501 |
+
"epoch": 0.5109849419896322,
|
| 1502 |
+
"grad_norm": 0.5444555426859482,
|
| 1503 |
+
"learning_rate": 5.678976463760635e-06,
|
| 1504 |
+
"loss": 1.2561,
|
| 1505 |
+
"step": 1035
|
| 1506 |
+
},
|
| 1507 |
+
{
|
| 1508 |
+
"epoch": 0.5134534682794372,
|
| 1509 |
+
"grad_norm": 0.5663913305474535,
|
| 1510 |
+
"learning_rate": 5.636244682367937e-06,
|
| 1511 |
+
"loss": 1.2324,
|
| 1512 |
+
"step": 1040
|
| 1513 |
+
},
|
| 1514 |
+
{
|
| 1515 |
+
"epoch": 0.5159219945692421,
|
| 1516 |
+
"grad_norm": 0.6001697304401695,
|
| 1517 |
+
"learning_rate": 5.593465611501127e-06,
|
| 1518 |
+
"loss": 1.2206,
|
| 1519 |
+
"step": 1045
|
| 1520 |
+
},
|
| 1521 |
+
{
|
| 1522 |
+
"epoch": 0.5183905208590471,
|
| 1523 |
+
"grad_norm": 0.5922209574486257,
|
| 1524 |
+
"learning_rate": 5.5506424307541895e-06,
|
| 1525 |
+
"loss": 1.2777,
|
| 1526 |
+
"step": 1050
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 0.5208590471488521,
|
| 1530 |
+
"grad_norm": 0.5810845811643376,
|
| 1531 |
+
"learning_rate": 5.507778322999615e-06,
|
| 1532 |
+
"loss": 1.2186,
|
| 1533 |
+
"step": 1055
|
| 1534 |
+
},
|
| 1535 |
+
{
|
| 1536 |
+
"epoch": 0.5233275734386571,
|
| 1537 |
+
"grad_norm": 0.5661815755139697,
|
| 1538 |
+
"learning_rate": 5.464876474151835e-06,
|
| 1539 |
+
"loss": 1.2465,
|
| 1540 |
+
"step": 1060
|
| 1541 |
+
},
|
| 1542 |
+
{
|
| 1543 |
+
"epoch": 0.5257960997284621,
|
| 1544 |
+
"grad_norm": 0.6016645517449551,
|
| 1545 |
+
"learning_rate": 5.421940072930415e-06,
|
| 1546 |
+
"loss": 1.2269,
|
| 1547 |
+
"step": 1065
|
| 1548 |
+
},
|
| 1549 |
+
{
|
| 1550 |
+
"epoch": 0.5282646260182671,
|
| 1551 |
+
"grad_norm": 0.6268744087157316,
|
| 1552 |
+
"learning_rate": 5.3789723106230675e-06,
|
| 1553 |
+
"loss": 1.2089,
|
| 1554 |
+
"step": 1070
|
| 1555 |
+
},
|
| 1556 |
+
{
|
| 1557 |
+
"epoch": 0.530733152308072,
|
| 1558 |
+
"grad_norm": 0.5374231313658383,
|
| 1559 |
+
"learning_rate": 5.3359763808484396e-06,
|
| 1560 |
+
"loss": 1.2371,
|
| 1561 |
+
"step": 1075
|
| 1562 |
+
},
|
| 1563 |
+
{
|
| 1564 |
+
"epoch": 0.533201678597877,
|
| 1565 |
+
"grad_norm": 0.5696825743006079,
|
| 1566 |
+
"learning_rate": 5.292955479318756e-06,
|
| 1567 |
+
"loss": 1.2288,
|
| 1568 |
+
"step": 1080
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 0.535670204887682,
|
| 1572 |
+
"grad_norm": 0.5474403893705062,
|
| 1573 |
+
"learning_rate": 5.249912803602287e-06,
|
| 1574 |
+
"loss": 1.2631,
|
| 1575 |
+
"step": 1085
|
| 1576 |
+
},
|
| 1577 |
+
{
|
| 1578 |
+
"epoch": 0.538138731177487,
|
| 1579 |
+
"grad_norm": 0.611438366860115,
|
| 1580 |
+
"learning_rate": 5.206851552885691e-06,
|
| 1581 |
+
"loss": 1.2395,
|
| 1582 |
+
"step": 1090
|
| 1583 |
+
},
|
| 1584 |
+
{
|
| 1585 |
+
"epoch": 0.540607257467292,
|
| 1586 |
+
"grad_norm": 0.6437738368971478,
|
| 1587 |
+
"learning_rate": 5.163774927736228e-06,
|
| 1588 |
+
"loss": 1.3132,
|
| 1589 |
+
"step": 1095
|
| 1590 |
+
},
|
| 1591 |
+
{
|
| 1592 |
+
"epoch": 0.543075783757097,
|
| 1593 |
+
"grad_norm": 0.5438676695949717,
|
| 1594 |
+
"learning_rate": 5.120686129863882e-06,
|
| 1595 |
+
"loss": 1.2807,
|
| 1596 |
+
"step": 1100
|
| 1597 |
+
},
|
| 1598 |
+
{
|
| 1599 |
+
"epoch": 0.545544310046902,
|
| 1600 |
+
"grad_norm": 0.6135072081701597,
|
| 1601 |
+
"learning_rate": 5.077588361883379e-06,
|
| 1602 |
+
"loss": 1.2239,
|
| 1603 |
+
"step": 1105
|
| 1604 |
+
},
|
| 1605 |
+
{
|
| 1606 |
+
"epoch": 0.5480128363367069,
|
| 1607 |
+
"grad_norm": 0.546701645842348,
|
| 1608 |
+
"learning_rate": 5.0344848270761635e-06,
|
| 1609 |
+
"loss": 1.2121,
|
| 1610 |
+
"step": 1110
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 0.5504813626265119,
|
| 1614 |
+
"grad_norm": 0.6153049309551597,
|
| 1615 |
+
"learning_rate": 4.9913787291523e-06,
|
| 1616 |
+
"loss": 1.2832,
|
| 1617 |
+
"step": 1115
|
| 1618 |
+
},
|
| 1619 |
+
{
|
| 1620 |
+
"epoch": 0.5529498889163169,
|
| 1621 |
+
"grad_norm": 0.6148368644966669,
|
| 1622 |
+
"learning_rate": 4.948273272012363e-06,
|
| 1623 |
+
"loss": 1.2536,
|
| 1624 |
+
"step": 1120
|
| 1625 |
+
},
|
| 1626 |
+
{
|
| 1627 |
+
"epoch": 0.5554184152061219,
|
| 1628 |
+
"grad_norm": 0.5911800001869699,
|
| 1629 |
+
"learning_rate": 4.905171659509294e-06,
|
| 1630 |
+
"loss": 1.2789,
|
| 1631 |
+
"step": 1125
|
| 1632 |
+
},
|
| 1633 |
+
{
|
| 1634 |
+
"epoch": 0.5578869414959269,
|
| 1635 |
+
"grad_norm": 0.5450128065258734,
|
| 1636 |
+
"learning_rate": 4.862077095210284e-06,
|
| 1637 |
+
"loss": 1.1595,
|
| 1638 |
+
"step": 1130
|
| 1639 |
+
},
|
| 1640 |
+
{
|
| 1641 |
+
"epoch": 0.5603554677857319,
|
| 1642 |
+
"grad_norm": 0.5629093671549396,
|
| 1643 |
+
"learning_rate": 4.818992782158658e-06,
|
| 1644 |
+
"loss": 1.2854,
|
| 1645 |
+
"step": 1135
|
| 1646 |
+
},
|
| 1647 |
+
{
|
| 1648 |
+
"epoch": 0.562823994075537,
|
| 1649 |
+
"grad_norm": 0.6634778146032412,
|
| 1650 |
+
"learning_rate": 4.775921922635806e-06,
|
| 1651 |
+
"loss": 1.2405,
|
| 1652 |
+
"step": 1140
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 0.5652925203653419,
|
| 1656 |
+
"grad_norm": 0.5439361692157106,
|
| 1657 |
+
"learning_rate": 4.732867717923174e-06,
|
| 1658 |
+
"loss": 1.265,
|
| 1659 |
+
"step": 1145
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 0.5677610466551469,
|
| 1663 |
+
"grad_norm": 0.5860651769650387,
|
| 1664 |
+
"learning_rate": 4.689833368064326e-06,
|
| 1665 |
+
"loss": 1.2511,
|
| 1666 |
+
"step": 1150
|
| 1667 |
+
},
|
| 1668 |
+
{
|
| 1669 |
+
"epoch": 0.5702295729449519,
|
| 1670 |
+
"grad_norm": 0.627265270599233,
|
| 1671 |
+
"learning_rate": 4.646822071627089e-06,
|
| 1672 |
+
"loss": 1.2813,
|
| 1673 |
+
"step": 1155
|
| 1674 |
+
},
|
| 1675 |
+
{
|
| 1676 |
+
"epoch": 0.5726980992347569,
|
| 1677 |
+
"grad_norm": 0.5634927900565491,
|
| 1678 |
+
"learning_rate": 4.603837025465829e-06,
|
| 1679 |
+
"loss": 1.22,
|
| 1680 |
+
"step": 1160
|
| 1681 |
+
},
|
| 1682 |
+
{
|
| 1683 |
+
"epoch": 0.5751666255245619,
|
| 1684 |
+
"grad_norm": 0.6482363315867818,
|
| 1685 |
+
"learning_rate": 4.560881424483833e-06,
|
| 1686 |
+
"loss": 1.3095,
|
| 1687 |
+
"step": 1165
|
| 1688 |
+
},
|
| 1689 |
+
{
|
| 1690 |
+
"epoch": 0.5776351518143669,
|
| 1691 |
+
"grad_norm": 0.4805380958857345,
|
| 1692 |
+
"learning_rate": 4.517958461395846e-06,
|
| 1693 |
+
"loss": 1.2737,
|
| 1694 |
+
"step": 1170
|
| 1695 |
+
},
|
| 1696 |
+
{
|
| 1697 |
+
"epoch": 0.5801036781041718,
|
| 1698 |
+
"grad_norm": 0.5854150858325277,
|
| 1699 |
+
"learning_rate": 4.475071326490781e-06,
|
| 1700 |
+
"loss": 1.2282,
|
| 1701 |
+
"step": 1175
|
| 1702 |
+
},
|
| 1703 |
+
{
|
| 1704 |
+
"epoch": 0.5825722043939768,
|
| 1705 |
+
"grad_norm": 0.554230131541799,
|
| 1706 |
+
"learning_rate": 4.432223207394577e-06,
|
| 1707 |
+
"loss": 1.178,
|
| 1708 |
+
"step": 1180
|
| 1709 |
+
},
|
| 1710 |
+
{
|
| 1711 |
+
"epoch": 0.5850407306837818,
|
| 1712 |
+
"grad_norm": 0.6930360615517788,
|
| 1713 |
+
"learning_rate": 4.389417288833292e-06,
|
| 1714 |
+
"loss": 1.2781,
|
| 1715 |
+
"step": 1185
|
| 1716 |
+
},
|
| 1717 |
+
{
|
| 1718 |
+
"epoch": 0.5875092569735868,
|
| 1719 |
+
"grad_norm": 0.6042088339838697,
|
| 1720 |
+
"learning_rate": 4.346656752396388e-06,
|
| 1721 |
+
"loss": 1.2813,
|
| 1722 |
+
"step": 1190
|
| 1723 |
+
},
|
| 1724 |
+
{
|
| 1725 |
+
"epoch": 0.5899777832633918,
|
| 1726 |
+
"grad_norm": 0.6280387565672664,
|
| 1727 |
+
"learning_rate": 4.303944776300262e-06,
|
| 1728 |
+
"loss": 1.2433,
|
| 1729 |
+
"step": 1195
|
| 1730 |
+
},
|
| 1731 |
+
{
|
| 1732 |
+
"epoch": 0.5924463095531968,
|
| 1733 |
+
"grad_norm": 0.5502891803034431,
|
| 1734 |
+
"learning_rate": 4.261284535152016e-06,
|
| 1735 |
+
"loss": 1.1556,
|
| 1736 |
+
"step": 1200
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"epoch": 0.5924463095531968,
|
| 1740 |
+
"eval_loss": 1.2148913145065308,
|
| 1741 |
+
"eval_runtime": 2558.7024,
|
| 1742 |
+
"eval_samples_per_second": 1.563,
|
| 1743 |
+
"eval_steps_per_second": 0.131,
|
| 1744 |
+
"step": 1200
|
| 1745 |
+
},
|
| 1746 |
+
{
|
| 1747 |
+
"epoch": 0.5949148358430018,
|
| 1748 |
+
"grad_norm": 0.5429417971755677,
|
| 1749 |
+
"learning_rate": 4.218679199713505e-06,
|
| 1750 |
+
"loss": 1.2398,
|
| 1751 |
+
"step": 1205
|
| 1752 |
+
},
|
| 1753 |
+
{
|
| 1754 |
+
"epoch": 0.5973833621328067,
|
| 1755 |
+
"grad_norm": 0.5573592415141271,
|
| 1756 |
+
"learning_rate": 4.176131936665669e-06,
|
| 1757 |
+
"loss": 1.2348,
|
| 1758 |
+
"step": 1210
|
| 1759 |
+
},
|
| 1760 |
+
{
|
| 1761 |
+
"epoch": 0.5998518884226117,
|
| 1762 |
+
"grad_norm": 0.5662130620287456,
|
| 1763 |
+
"learning_rate": 4.133645908373159e-06,
|
| 1764 |
+
"loss": 1.1894,
|
| 1765 |
+
"step": 1215
|
| 1766 |
+
},
|
| 1767 |
+
{
|
| 1768 |
+
"epoch": 0.6023204147124167,
|
| 1769 |
+
"grad_norm": 0.5330337777111593,
|
| 1770 |
+
"learning_rate": 4.0912242726493e-06,
|
| 1771 |
+
"loss": 1.267,
|
| 1772 |
+
"step": 1220
|
| 1773 |
+
},
|
| 1774 |
+
{
|
| 1775 |
+
"epoch": 0.6047889410022217,
|
| 1776 |
+
"grad_norm": 0.589763462299109,
|
| 1777 |
+
"learning_rate": 4.048870182521374e-06,
|
| 1778 |
+
"loss": 1.2461,
|
| 1779 |
+
"step": 1225
|
| 1780 |
+
},
|
| 1781 |
+
{
|
| 1782 |
+
"epoch": 0.6072574672920267,
|
| 1783 |
+
"grad_norm": 0.5798241574940401,
|
| 1784 |
+
"learning_rate": 4.006586785996285e-06,
|
| 1785 |
+
"loss": 1.2503,
|
| 1786 |
+
"step": 1230
|
| 1787 |
+
},
|
| 1788 |
+
{
|
| 1789 |
+
"epoch": 0.6097259935818317,
|
| 1790 |
+
"grad_norm": 0.5714021679563045,
|
| 1791 |
+
"learning_rate": 3.96437722582656e-06,
|
| 1792 |
+
"loss": 1.2322,
|
| 1793 |
+
"step": 1235
|
| 1794 |
+
},
|
| 1795 |
+
{
|
| 1796 |
+
"epoch": 0.6121945198716366,
|
| 1797 |
+
"grad_norm": 0.5926307509257247,
|
| 1798 |
+
"learning_rate": 3.922244639276773e-06,
|
| 1799 |
+
"loss": 1.2692,
|
| 1800 |
+
"step": 1240
|
| 1801 |
+
},
|
| 1802 |
+
{
|
| 1803 |
+
"epoch": 0.6146630461614416,
|
| 1804 |
+
"grad_norm": 0.6016557090563102,
|
| 1805 |
+
"learning_rate": 3.880192157890365e-06,
|
| 1806 |
+
"loss": 1.2642,
|
| 1807 |
+
"step": 1245
|
| 1808 |
+
},
|
| 1809 |
+
{
|
| 1810 |
+
"epoch": 0.6171315724512466,
|
| 1811 |
+
"grad_norm": 0.5454381088492659,
|
| 1812 |
+
"learning_rate": 3.838222907256884e-06,
|
| 1813 |
+
"loss": 1.239,
|
| 1814 |
+
"step": 1250
|
| 1815 |
+
},
|
| 1816 |
+
{
|
| 1817 |
+
"epoch": 0.6196000987410516,
|
| 1818 |
+
"grad_norm": 0.5582749852816064,
|
| 1819 |
+
"learning_rate": 3.7963400067796774e-06,
|
| 1820 |
+
"loss": 1.2851,
|
| 1821 |
+
"step": 1255
|
| 1822 |
+
},
|
| 1823 |
+
{
|
| 1824 |
+
"epoch": 0.6220686250308566,
|
| 1825 |
+
"grad_norm": 0.5562967849735465,
|
| 1826 |
+
"learning_rate": 3.7545465694440363e-06,
|
| 1827 |
+
"loss": 1.2432,
|
| 1828 |
+
"step": 1260
|
| 1829 |
+
},
|
| 1830 |
+
{
|
| 1831 |
+
"epoch": 0.6245371513206616,
|
| 1832 |
+
"grad_norm": 0.5419669962437569,
|
| 1833 |
+
"learning_rate": 3.7128457015858198e-06,
|
| 1834 |
+
"loss": 1.2103,
|
| 1835 |
+
"step": 1265
|
| 1836 |
+
},
|
| 1837 |
+
{
|
| 1838 |
+
"epoch": 0.6270056776104665,
|
| 1839 |
+
"grad_norm": 0.558873424565738,
|
| 1840 |
+
"learning_rate": 3.6712405026605792e-06,
|
| 1841 |
+
"loss": 1.2388,
|
| 1842 |
+
"step": 1270
|
| 1843 |
+
},
|
| 1844 |
+
{
|
| 1845 |
+
"epoch": 0.6294742039002715,
|
| 1846 |
+
"grad_norm": 0.5712282397945332,
|
| 1847 |
+
"learning_rate": 3.6297340650131785e-06,
|
| 1848 |
+
"loss": 1.2819,
|
| 1849 |
+
"step": 1275
|
| 1850 |
+
},
|
| 1851 |
+
{
|
| 1852 |
+
"epoch": 0.6319427301900765,
|
| 1853 |
+
"grad_norm": 0.5643697726223241,
|
| 1854 |
+
"learning_rate": 3.5883294736479612e-06,
|
| 1855 |
+
"loss": 1.2386,
|
| 1856 |
+
"step": 1280
|
| 1857 |
+
},
|
| 1858 |
+
{
|
| 1859 |
+
"epoch": 0.6344112564798815,
|
| 1860 |
+
"grad_norm": 0.6332020317807455,
|
| 1861 |
+
"learning_rate": 3.5470298059994545e-06,
|
| 1862 |
+
"loss": 1.2677,
|
| 1863 |
+
"step": 1285
|
| 1864 |
+
},
|
| 1865 |
+
{
|
| 1866 |
+
"epoch": 0.6368797827696865,
|
| 1867 |
+
"grad_norm": 0.6276157822500693,
|
| 1868 |
+
"learning_rate": 3.5058381317036285e-06,
|
| 1869 |
+
"loss": 1.2137,
|
| 1870 |
+
"step": 1290
|
| 1871 |
+
},
|
| 1872 |
+
{
|
| 1873 |
+
"epoch": 0.6393483090594915,
|
| 1874 |
+
"grad_norm": 0.5139753708360036,
|
| 1875 |
+
"learning_rate": 3.46475751236975e-06,
|
| 1876 |
+
"loss": 1.2436,
|
| 1877 |
+
"step": 1295
|
| 1878 |
+
},
|
| 1879 |
+
{
|
| 1880 |
+
"epoch": 0.6418168353492965,
|
| 1881 |
+
"grad_norm": 0.5868933304811402,
|
| 1882 |
+
"learning_rate": 3.423791001352823e-06,
|
| 1883 |
+
"loss": 1.1681,
|
| 1884 |
+
"step": 1300
|
| 1885 |
+
},
|
| 1886 |
+
{
|
| 1887 |
+
"epoch": 0.6442853616391014,
|
| 1888 |
+
"grad_norm": 0.5592137564928078,
|
| 1889 |
+
"learning_rate": 3.382941643526644e-06,
|
| 1890 |
+
"loss": 1.2443,
|
| 1891 |
+
"step": 1305
|
| 1892 |
+
},
|
| 1893 |
+
{
|
| 1894 |
+
"epoch": 0.6467538879289064,
|
| 1895 |
+
"grad_norm": 0.567548616583169,
|
| 1896 |
+
"learning_rate": 3.3422124750574902e-06,
|
| 1897 |
+
"loss": 1.2604,
|
| 1898 |
+
"step": 1310
|
| 1899 |
+
},
|
| 1900 |
+
{
|
| 1901 |
+
"epoch": 0.6492224142187114,
|
| 1902 |
+
"grad_norm": 0.568882999500645,
|
| 1903 |
+
"learning_rate": 3.3016065231784587e-06,
|
| 1904 |
+
"loss": 1.1595,
|
| 1905 |
+
"step": 1315
|
| 1906 |
+
},
|
| 1907 |
+
{
|
| 1908 |
+
"epoch": 0.6516909405085164,
|
| 1909 |
+
"grad_norm": 0.628304707671549,
|
| 1910 |
+
"learning_rate": 3.2611268059644535e-06,
|
| 1911 |
+
"loss": 1.2841,
|
| 1912 |
+
"step": 1320
|
| 1913 |
+
},
|
| 1914 |
+
{
|
| 1915 |
+
"epoch": 0.6541594667983214,
|
| 1916 |
+
"grad_norm": 0.5686219665932154,
|
| 1917 |
+
"learning_rate": 3.2207763321078737e-06,
|
| 1918 |
+
"loss": 1.2347,
|
| 1919 |
+
"step": 1325
|
| 1920 |
+
},
|
| 1921 |
+
{
|
| 1922 |
+
"epoch": 0.6566279930881264,
|
| 1923 |
+
"grad_norm": 0.6424587872522304,
|
| 1924 |
+
"learning_rate": 3.1805581006949856e-06,
|
| 1925 |
+
"loss": 1.2329,
|
| 1926 |
+
"step": 1330
|
| 1927 |
+
},
|
| 1928 |
+
{
|
| 1929 |
+
"epoch": 0.6590965193779313,
|
| 1930 |
+
"grad_norm": 0.6654374856920555,
|
| 1931 |
+
"learning_rate": 3.1404751009830124e-06,
|
| 1932 |
+
"loss": 1.2423,
|
| 1933 |
+
"step": 1335
|
| 1934 |
+
},
|
| 1935 |
+
{
|
| 1936 |
+
"epoch": 0.6615650456677363,
|
| 1937 |
+
"grad_norm": 0.5206675422652753,
|
| 1938 |
+
"learning_rate": 3.100530312177956e-06,
|
| 1939 |
+
"loss": 1.2329,
|
| 1940 |
+
"step": 1340
|
| 1941 |
+
},
|
| 1942 |
+
{
|
| 1943 |
+
"epoch": 0.6640335719575413,
|
| 1944 |
+
"grad_norm": 0.6656795155578475,
|
| 1945 |
+
"learning_rate": 3.0607267032131704e-06,
|
| 1946 |
+
"loss": 1.3062,
|
| 1947 |
+
"step": 1345
|
| 1948 |
+
},
|
| 1949 |
+
{
|
| 1950 |
+
"epoch": 0.6665020982473463,
|
| 1951 |
+
"grad_norm": 0.6071844948708964,
|
| 1952 |
+
"learning_rate": 3.0210672325286806e-06,
|
| 1953 |
+
"loss": 1.2656,
|
| 1954 |
+
"step": 1350
|
| 1955 |
+
},
|
| 1956 |
+
{
|
| 1957 |
+
"epoch": 0.6689706245371513,
|
| 1958 |
+
"grad_norm": 0.6211025479318184,
|
| 1959 |
+
"learning_rate": 2.9815548478513034e-06,
|
| 1960 |
+
"loss": 1.2167,
|
| 1961 |
+
"step": 1355
|
| 1962 |
+
},
|
| 1963 |
+
{
|
| 1964 |
+
"epoch": 0.6714391508269563,
|
| 1965 |
+
"grad_norm": 0.5801456765244887,
|
| 1966 |
+
"learning_rate": 2.9421924859755525e-06,
|
| 1967 |
+
"loss": 1.2249,
|
| 1968 |
+
"step": 1360
|
| 1969 |
+
},
|
| 1970 |
+
{
|
| 1971 |
+
"epoch": 0.6739076771167613,
|
| 1972 |
+
"grad_norm": 0.564862030285346,
|
| 1973 |
+
"learning_rate": 2.9029830725453545e-06,
|
| 1974 |
+
"loss": 1.2414,
|
| 1975 |
+
"step": 1365
|
| 1976 |
+
},
|
| 1977 |
+
{
|
| 1978 |
+
"epoch": 0.6763762034065662,
|
| 1979 |
+
"grad_norm": 0.5538133203567932,
|
| 1980 |
+
"learning_rate": 2.8639295218366115e-06,
|
| 1981 |
+
"loss": 1.2191,
|
| 1982 |
+
"step": 1370
|
| 1983 |
+
},
|
| 1984 |
+
{
|
| 1985 |
+
"epoch": 0.6788447296963712,
|
| 1986 |
+
"grad_norm": 0.5925104037633543,
|
| 1987 |
+
"learning_rate": 2.8250347365405737e-06,
|
| 1988 |
+
"loss": 1.2318,
|
| 1989 |
+
"step": 1375
|
| 1990 |
+
},
|
| 1991 |
+
{
|
| 1992 |
+
"epoch": 0.6813132559861762,
|
| 1993 |
+
"grad_norm": 0.6173909875052214,
|
| 1994 |
+
"learning_rate": 2.78630160754811e-06,
|
| 1995 |
+
"loss": 1.2555,
|
| 1996 |
+
"step": 1380
|
| 1997 |
+
},
|
| 1998 |
+
{
|
| 1999 |
+
"epoch": 0.6837817822759812,
|
| 2000 |
+
"grad_norm": 0.6579800769123958,
|
| 2001 |
+
"learning_rate": 2.747733013734835e-06,
|
| 2002 |
+
"loss": 1.2553,
|
| 2003 |
+
"step": 1385
|
| 2004 |
+
},
|
| 2005 |
+
{
|
| 2006 |
+
"epoch": 0.6862503085657862,
|
| 2007 |
+
"grad_norm": 0.6097488788659552,
|
| 2008 |
+
"learning_rate": 2.709331821747133e-06,
|
| 2009 |
+
"loss": 1.2482,
|
| 2010 |
+
"step": 1390
|
| 2011 |
+
},
|
| 2012 |
+
{
|
| 2013 |
+
"epoch": 0.6887188348555913,
|
| 2014 |
+
"grad_norm": 0.5717544066297715,
|
| 2015 |
+
"learning_rate": 2.6711008857890928e-06,
|
| 2016 |
+
"loss": 1.2477,
|
| 2017 |
+
"step": 1395
|
| 2018 |
+
},
|
| 2019 |
+
{
|
| 2020 |
+
"epoch": 0.6911873611453963,
|
| 2021 |
+
"grad_norm": 0.5675063300875494,
|
| 2022 |
+
"learning_rate": 2.63304304741037e-06,
|
| 2023 |
+
"loss": 1.2386,
|
| 2024 |
+
"step": 1400
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"epoch": 0.6911873611453963,
|
| 2028 |
+
"eval_loss": 1.211606740951538,
|
| 2029 |
+
"eval_runtime": 2914.6181,
|
| 2030 |
+
"eval_samples_per_second": 1.372,
|
| 2031 |
+
"eval_steps_per_second": 0.115,
|
| 2032 |
+
"step": 1400
|
| 2033 |
+
},
|
| 2034 |
+
{
|
| 2035 |
+
"epoch": 0.6936558874352012,
|
| 2036 |
+
"grad_norm": 0.623871781326139,
|
| 2037 |
+
"learning_rate": 2.595161135294978e-06,
|
| 2038 |
+
"loss": 1.2484,
|
| 2039 |
+
"step": 1405
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 0.6961244137250062,
|
| 2043 |
+
"grad_norm": 0.5967791678571923,
|
| 2044 |
+
"learning_rate": 2.55745796505105e-06,
|
| 2045 |
+
"loss": 1.2816,
|
| 2046 |
+
"step": 1410
|
| 2047 |
+
},
|
| 2048 |
+
{
|
| 2049 |
+
"epoch": 0.6985929400148112,
|
| 2050 |
+
"grad_norm": 0.5958918786737188,
|
| 2051 |
+
"learning_rate": 2.5199363390015645e-06,
|
| 2052 |
+
"loss": 1.2518,
|
| 2053 |
+
"step": 1415
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"epoch": 0.7010614663046162,
|
| 2057 |
+
"grad_norm": 0.5716469845277612,
|
| 2058 |
+
"learning_rate": 2.482599045976059e-06,
|
| 2059 |
+
"loss": 1.2518,
|
| 2060 |
+
"step": 1420
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 0.7035299925944212,
|
| 2064 |
+
"grad_norm": 0.5601354887821722,
|
| 2065 |
+
"learning_rate": 2.445448861103348e-06,
|
| 2066 |
+
"loss": 1.2114,
|
| 2067 |
+
"step": 1425
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 0.7059985188842262,
|
| 2071 |
+
"grad_norm": 0.5783618487395104,
|
| 2072 |
+
"learning_rate": 2.408488545605265e-06,
|
| 2073 |
+
"loss": 1.2801,
|
| 2074 |
+
"step": 1430
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 0.7084670451740311,
|
| 2078 |
+
"grad_norm": 0.600120666255256,
|
| 2079 |
+
"learning_rate": 2.3717208465914193e-06,
|
| 2080 |
+
"loss": 1.2928,
|
| 2081 |
+
"step": 1435
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 0.7109355714638361,
|
| 2085 |
+
"grad_norm": 0.6823362059514299,
|
| 2086 |
+
"learning_rate": 2.3351484968550264e-06,
|
| 2087 |
+
"loss": 1.2306,
|
| 2088 |
+
"step": 1440
|
| 2089 |
+
},
|
| 2090 |
+
{
|
| 2091 |
+
"epoch": 0.7134040977536411,
|
| 2092 |
+
"grad_norm": 0.5869728269343567,
|
| 2093 |
+
"learning_rate": 2.298774214669785e-06,
|
| 2094 |
+
"loss": 1.2417,
|
| 2095 |
+
"step": 1445
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 0.7158726240434461,
|
| 2099 |
+
"grad_norm": 0.597629982893601,
|
| 2100 |
+
"learning_rate": 2.2626007035878377e-06,
|
| 2101 |
+
"loss": 1.1912,
|
| 2102 |
+
"step": 1450
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 0.7183411503332511,
|
| 2106 |
+
"grad_norm": 0.6222473980576229,
|
| 2107 |
+
"learning_rate": 2.226630652238836e-06,
|
| 2108 |
+
"loss": 1.2083,
|
| 2109 |
+
"step": 1455
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 0.7208096766230561,
|
| 2113 |
+
"grad_norm": 0.5978767327421509,
|
| 2114 |
+
"learning_rate": 2.1908667341300923e-06,
|
| 2115 |
+
"loss": 1.2577,
|
| 2116 |
+
"step": 1460
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"epoch": 0.723278202912861,
|
| 2120 |
+
"grad_norm": 0.6156905912164004,
|
| 2121 |
+
"learning_rate": 2.155311607447877e-06,
|
| 2122 |
+
"loss": 1.2922,
|
| 2123 |
+
"step": 1465
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 0.725746729202666,
|
| 2127 |
+
"grad_norm": 0.6341472520929511,
|
| 2128 |
+
"learning_rate": 2.1199679148598434e-06,
|
| 2129 |
+
"loss": 1.2667,
|
| 2130 |
+
"step": 1470
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 0.728215255492471,
|
| 2134 |
+
"grad_norm": 0.5655996654676207,
|
| 2135 |
+
"learning_rate": 2.084838283318616e-06,
|
| 2136 |
+
"loss": 1.1939,
|
| 2137 |
+
"step": 1475
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 0.730683781782276,
|
| 2141 |
+
"grad_norm": 0.5824088027115487,
|
| 2142 |
+
"learning_rate": 2.0499253238665284e-06,
|
| 2143 |
+
"loss": 1.242,
|
| 2144 |
+
"step": 1480
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 0.733152308072081,
|
| 2148 |
+
"grad_norm": 0.6063388402546945,
|
| 2149 |
+
"learning_rate": 2.0152316314415602e-06,
|
| 2150 |
+
"loss": 1.2482,
|
| 2151 |
+
"step": 1485
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 0.735620834361886,
|
| 2155 |
+
"grad_norm": 0.6226805122487513,
|
| 2156 |
+
"learning_rate": 1.9807597846844737e-06,
|
| 2157 |
+
"loss": 1.255,
|
| 2158 |
+
"step": 1490
|
| 2159 |
+
},
|
| 2160 |
+
{
|
| 2161 |
+
"epoch": 0.738089360651691,
|
| 2162 |
+
"grad_norm": 0.5854379294811827,
|
| 2163 |
+
"learning_rate": 1.9465123457471395e-06,
|
| 2164 |
+
"loss": 1.1786,
|
| 2165 |
+
"step": 1495
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 0.7405578869414959,
|
| 2169 |
+
"grad_norm": 0.5577052246580572,
|
| 2170 |
+
"learning_rate": 1.9124918601021124e-06,
|
| 2171 |
+
"loss": 1.2358,
|
| 2172 |
+
"step": 1500
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 0.7430264132313009,
|
| 2176 |
+
"grad_norm": 0.5754079743445688,
|
| 2177 |
+
"learning_rate": 1.8787008563534326e-06,
|
| 2178 |
+
"loss": 1.1945,
|
| 2179 |
+
"step": 1505
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 0.7454949395211059,
|
| 2183 |
+
"grad_norm": 0.6099556355269008,
|
| 2184 |
+
"learning_rate": 1.845141846048691e-06,
|
| 2185 |
+
"loss": 1.2379,
|
| 2186 |
+
"step": 1510
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 0.7479634658109109,
|
| 2190 |
+
"grad_norm": 0.5782704010521243,
|
| 2191 |
+
"learning_rate": 1.8118173234923447e-06,
|
| 2192 |
+
"loss": 1.2542,
|
| 2193 |
+
"step": 1515
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 0.7504319921007159,
|
| 2197 |
+
"grad_norm": 0.5382858254483444,
|
| 2198 |
+
"learning_rate": 1.778729765560337e-06,
|
| 2199 |
+
"loss": 1.2327,
|
| 2200 |
+
"step": 1520
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 0.7529005183905209,
|
| 2204 |
+
"grad_norm": 0.6082642317550977,
|
| 2205 |
+
"learning_rate": 1.7458816315159937e-06,
|
| 2206 |
+
"loss": 1.2631,
|
| 2207 |
+
"step": 1525
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 0.7553690446803258,
|
| 2211 |
+
"grad_norm": 0.6120502232540203,
|
| 2212 |
+
"learning_rate": 1.7132753628272403e-06,
|
| 2213 |
+
"loss": 1.2687,
|
| 2214 |
+
"step": 1530
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 0.7578375709701308,
|
| 2218 |
+
"grad_norm": 0.5800190917782422,
|
| 2219 |
+
"learning_rate": 1.6809133829851344e-06,
|
| 2220 |
+
"loss": 1.1809,
|
| 2221 |
+
"step": 1535
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 0.7603060972599358,
|
| 2225 |
+
"grad_norm": 0.6248767795672576,
|
| 2226 |
+
"learning_rate": 1.6487980973237434e-06,
|
| 2227 |
+
"loss": 1.2102,
|
| 2228 |
+
"step": 1540
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 0.7627746235497408,
|
| 2232 |
+
"grad_norm": 0.6214869106372124,
|
| 2233 |
+
"learning_rate": 1.6169318928413574e-06,
|
| 2234 |
+
"loss": 1.3183,
|
| 2235 |
+
"step": 1545
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 0.7652431498395458,
|
| 2239 |
+
"grad_norm": 0.6509287986960063,
|
| 2240 |
+
"learning_rate": 1.5853171380230791e-06,
|
| 2241 |
+
"loss": 1.2394,
|
| 2242 |
+
"step": 1550
|
| 2243 |
+
},
|
| 2244 |
+
{
|
| 2245 |
+
"epoch": 0.7677116761293508,
|
| 2246 |
+
"grad_norm": 0.5548564286839581,
|
| 2247 |
+
"learning_rate": 1.5539561826647832e-06,
|
| 2248 |
+
"loss": 1.2278,
|
| 2249 |
+
"step": 1555
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 0.7701802024191557,
|
| 2253 |
+
"grad_norm": 0.5873399173100068,
|
| 2254 |
+
"learning_rate": 1.5228513576984633e-06,
|
| 2255 |
+
"loss": 1.2419,
|
| 2256 |
+
"step": 1560
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 0.7726487287089607,
|
| 2260 |
+
"grad_norm": 0.5698526241039991,
|
| 2261 |
+
"learning_rate": 1.4920049750189852e-06,
|
| 2262 |
+
"loss": 1.2134,
|
| 2263 |
+
"step": 1565
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 0.7751172549987657,
|
| 2267 |
+
"grad_norm": 0.5462525752885333,
|
| 2268 |
+
"learning_rate": 1.4614193273122562e-06,
|
| 2269 |
+
"loss": 1.2013,
|
| 2270 |
+
"step": 1570
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 0.7775857812885707,
|
| 2274 |
+
"grad_norm": 0.5604406125512932,
|
| 2275 |
+
"learning_rate": 1.4310966878848116e-06,
|
| 2276 |
+
"loss": 1.2319,
|
| 2277 |
+
"step": 1575
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 0.7800543075783757,
|
| 2281 |
+
"grad_norm": 0.5512496837811336,
|
| 2282 |
+
"learning_rate": 1.401039310494855e-06,
|
| 2283 |
+
"loss": 1.2436,
|
| 2284 |
+
"step": 1580
|
| 2285 |
+
},
|
| 2286 |
+
{
|
| 2287 |
+
"epoch": 0.7825228338681807,
|
| 2288 |
+
"grad_norm": 0.6804998312407946,
|
| 2289 |
+
"learning_rate": 1.3712494291847416e-06,
|
| 2290 |
+
"loss": 1.2567,
|
| 2291 |
+
"step": 1585
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 0.7849913601579857,
|
| 2295 |
+
"grad_norm": 0.6655723000722049,
|
| 2296 |
+
"learning_rate": 1.3417292581149388e-06,
|
| 2297 |
+
"loss": 1.2682,
|
| 2298 |
+
"step": 1590
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 0.7874598864477906,
|
| 2302 |
+
"grad_norm": 0.539222744257867,
|
| 2303 |
+
"learning_rate": 1.3124809913994458e-06,
|
| 2304 |
+
"loss": 1.2009,
|
| 2305 |
+
"step": 1595
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 0.7899284127375956,
|
| 2309 |
+
"grad_norm": 0.622721298212167,
|
| 2310 |
+
"learning_rate": 1.2835068029427188e-06,
|
| 2311 |
+
"loss": 1.2661,
|
| 2312 |
+
"step": 1600
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 0.7899284127375956,
|
| 2316 |
+
"eval_loss": 1.2096235752105713,
|
| 2317 |
+
"eval_runtime": 2576.8943,
|
| 2318 |
+
"eval_samples_per_second": 1.552,
|
| 2319 |
+
"eval_steps_per_second": 0.13,
|
| 2320 |
+
"step": 1600
|
| 2321 |
+
},
|
| 2322 |
+
{
|
| 2323 |
+
"epoch": 0.7923969390274006,
|
| 2324 |
+
"grad_norm": 0.5470842930259888,
|
| 2325 |
+
"learning_rate": 1.2548088462781006e-06,
|
| 2326 |
+
"loss": 1.2244,
|
| 2327 |
+
"step": 1605
|
| 2328 |
+
},
|
| 2329 |
+
{
|
| 2330 |
+
"epoch": 0.7948654653172056,
|
| 2331 |
+
"grad_norm": 0.5718801309412294,
|
| 2332 |
+
"learning_rate": 1.2263892544077439e-06,
|
| 2333 |
+
"loss": 1.2498,
|
| 2334 |
+
"step": 1610
|
| 2335 |
+
},
|
| 2336 |
+
{
|
| 2337 |
+
"epoch": 0.7973339916070106,
|
| 2338 |
+
"grad_norm": 0.5818869817428877,
|
| 2339 |
+
"learning_rate": 1.1982501396440831e-06,
|
| 2340 |
+
"loss": 1.2044,
|
| 2341 |
+
"step": 1615
|
| 2342 |
+
},
|
| 2343 |
+
{
|
| 2344 |
+
"epoch": 0.7998025178968156,
|
| 2345 |
+
"grad_norm": 0.5534354350847027,
|
| 2346 |
+
"learning_rate": 1.1703935934528327e-06,
|
| 2347 |
+
"loss": 1.2328,
|
| 2348 |
+
"step": 1620
|
| 2349 |
+
},
|
| 2350 |
+
{
|
| 2351 |
+
"epoch": 0.8022710441866205,
|
| 2352 |
+
"grad_norm": 0.5862274808604895,
|
| 2353 |
+
"learning_rate": 1.1428216862975383e-06,
|
| 2354 |
+
"loss": 1.2741,
|
| 2355 |
+
"step": 1625
|
| 2356 |
+
},
|
| 2357 |
+
{
|
| 2358 |
+
"epoch": 0.8047395704764255,
|
| 2359 |
+
"grad_norm": 0.5781950796979888,
|
| 2360 |
+
"learning_rate": 1.1155364674856834e-06,
|
| 2361 |
+
"loss": 1.2679,
|
| 2362 |
+
"step": 1630
|
| 2363 |
+
},
|
| 2364 |
+
{
|
| 2365 |
+
"epoch": 0.8072080967662305,
|
| 2366 |
+
"grad_norm": 0.5751302301159884,
|
| 2367 |
+
"learning_rate": 1.088539965016377e-06,
|
| 2368 |
+
"loss": 1.2153,
|
| 2369 |
+
"step": 1635
|
| 2370 |
+
},
|
| 2371 |
+
{
|
| 2372 |
+
"epoch": 0.8096766230560355,
|
| 2373 |
+
"grad_norm": 0.6150065644184977,
|
| 2374 |
+
"learning_rate": 1.0618341854296176e-06,
|
| 2375 |
+
"loss": 1.2245,
|
| 2376 |
+
"step": 1640
|
| 2377 |
+
},
|
| 2378 |
+
{
|
| 2379 |
+
"epoch": 0.8121451493458405,
|
| 2380 |
+
"grad_norm": 0.5893743060234344,
|
| 2381 |
+
"learning_rate": 1.0354211136571586e-06,
|
| 2382 |
+
"loss": 1.2091,
|
| 2383 |
+
"step": 1645
|
| 2384 |
+
},
|
| 2385 |
+
{
|
| 2386 |
+
"epoch": 0.8146136756356456,
|
| 2387 |
+
"grad_norm": 0.554001627193442,
|
| 2388 |
+
"learning_rate": 1.0093027128749722e-06,
|
| 2389 |
+
"loss": 1.22,
|
| 2390 |
+
"step": 1650
|
| 2391 |
+
},
|
| 2392 |
+
{
|
| 2393 |
+
"epoch": 0.8170822019254506,
|
| 2394 |
+
"grad_norm": 0.5554016650617593,
|
| 2395 |
+
"learning_rate": 9.834809243573406e-07,
|
| 2396 |
+
"loss": 1.2736,
|
| 2397 |
+
"step": 1655
|
| 2398 |
+
},
|
| 2399 |
+
{
|
| 2400 |
+
"epoch": 0.8195507282152555,
|
| 2401 |
+
"grad_norm": 0.6467820952863279,
|
| 2402 |
+
"learning_rate": 9.57957667332562e-07,
|
| 2403 |
+
"loss": 1.2504,
|
| 2404 |
+
"step": 1660
|
| 2405 |
+
},
|
| 2406 |
+
{
|
| 2407 |
+
"epoch": 0.8220192545050605,
|
| 2408 |
+
"grad_norm": 0.5388841867240308,
|
| 2409 |
+
"learning_rate": 9.327348388403063e-07,
|
| 2410 |
+
"loss": 1.2134,
|
| 2411 |
+
"step": 1665
|
| 2412 |
+
},
|
| 2413 |
+
{
|
| 2414 |
+
"epoch": 0.8244877807948655,
|
| 2415 |
+
"grad_norm": 0.5511949198965124,
|
| 2416 |
+
"learning_rate": 9.078143135906154e-07,
|
| 2417 |
+
"loss": 1.2373,
|
| 2418 |
+
"step": 1670
|
| 2419 |
+
},
|
| 2420 |
+
{
|
| 2421 |
+
"epoch": 0.8269563070846705,
|
| 2422 |
+
"grad_norm": 0.5662492648467455,
|
| 2423 |
+
"learning_rate": 8.831979438245619e-07,
|
| 2424 |
+
"loss": 1.2379,
|
| 2425 |
+
"step": 1675
|
| 2426 |
+
},
|
| 2427 |
+
{
|
| 2428 |
+
"epoch": 0.8294248333744755,
|
| 2429 |
+
"grad_norm": 0.6308948625824087,
|
| 2430 |
+
"learning_rate": 8.588875591765838e-07,
|
| 2431 |
+
"loss": 1.1868,
|
| 2432 |
+
"step": 1680
|
| 2433 |
+
},
|
| 2434 |
+
{
|
| 2435 |
+
"epoch": 0.8318933596642805,
|
| 2436 |
+
"grad_norm": 0.576660126030343,
|
| 2437 |
+
"learning_rate": 8.348849665384906e-07,
|
| 2438 |
+
"loss": 1.2891,
|
| 2439 |
+
"step": 1685
|
| 2440 |
+
},
|
| 2441 |
+
{
|
| 2442 |
+
"epoch": 0.8343618859540854,
|
| 2443 |
+
"grad_norm": 0.556606789107177,
|
| 2444 |
+
"learning_rate": 8.111919499251653e-07,
|
| 2445 |
+
"loss": 1.2021,
|
| 2446 |
+
"step": 1690
|
| 2447 |
+
},
|
| 2448 |
+
{
|
| 2449 |
+
"epoch": 0.8368304122438904,
|
| 2450 |
+
"grad_norm": 0.5661740275037651,
|
| 2451 |
+
"learning_rate": 7.878102703419683e-07,
|
| 2452 |
+
"loss": 1.2536,
|
| 2453 |
+
"step": 1695
|
| 2454 |
+
},
|
| 2455 |
+
{
|
| 2456 |
+
"epoch": 0.8392989385336954,
|
| 2457 |
+
"grad_norm": 0.5967205392911274,
|
| 2458 |
+
"learning_rate": 7.647416656538464e-07,
|
| 2459 |
+
"loss": 1.2373,
|
| 2460 |
+
"step": 1700
|
| 2461 |
+
},
|
| 2462 |
+
{
|
| 2463 |
+
"epoch": 0.8417674648235004,
|
| 2464 |
+
"grad_norm": 0.5528061162446166,
|
| 2465 |
+
"learning_rate": 7.419878504561651e-07,
|
| 2466 |
+
"loss": 1.2199,
|
| 2467 |
+
"step": 1705
|
| 2468 |
+
},
|
| 2469 |
+
{
|
| 2470 |
+
"epoch": 0.8442359911133054,
|
| 2471 |
+
"grad_norm": 0.6479872928308008,
|
| 2472 |
+
"learning_rate": 7.195505159472726e-07,
|
| 2473 |
+
"loss": 1.2368,
|
| 2474 |
+
"step": 1710
|
| 2475 |
+
},
|
| 2476 |
+
{
|
| 2477 |
+
"epoch": 0.8467045174031104,
|
| 2478 |
+
"grad_norm": 0.594834011459554,
|
| 2479 |
+
"learning_rate": 6.974313298027946e-07,
|
| 2480 |
+
"loss": 1.1997,
|
| 2481 |
+
"step": 1715
|
| 2482 |
+
},
|
| 2483 |
+
{
|
| 2484 |
+
"epoch": 0.8491730436929154,
|
| 2485 |
+
"grad_norm": 0.5442970599231537,
|
| 2486 |
+
"learning_rate": 6.756319360516856e-07,
|
| 2487 |
+
"loss": 1.2037,
|
| 2488 |
+
"step": 1720
|
| 2489 |
+
},
|
| 2490 |
+
{
|
| 2491 |
+
"epoch": 0.8516415699827203,
|
| 2492 |
+
"grad_norm": 0.6655980946948994,
|
| 2493 |
+
"learning_rate": 6.541539549540383e-07,
|
| 2494 |
+
"loss": 1.3013,
|
| 2495 |
+
"step": 1725
|
| 2496 |
+
},
|
| 2497 |
+
{
|
| 2498 |
+
"epoch": 0.8541100962725253,
|
| 2499 |
+
"grad_norm": 0.599651741019629,
|
| 2500 |
+
"learning_rate": 6.329989828806482e-07,
|
| 2501 |
+
"loss": 1.2454,
|
| 2502 |
+
"step": 1730
|
| 2503 |
+
},
|
| 2504 |
+
{
|
| 2505 |
+
"epoch": 0.8565786225623303,
|
| 2506 |
+
"grad_norm": 0.7507415296204425,
|
| 2507 |
+
"learning_rate": 6.121685921943688e-07,
|
| 2508 |
+
"loss": 1.2347,
|
| 2509 |
+
"step": 1735
|
| 2510 |
+
},
|
| 2511 |
+
{
|
| 2512 |
+
"epoch": 0.8590471488521353,
|
| 2513 |
+
"grad_norm": 0.5883088948787556,
|
| 2514 |
+
"learning_rate": 5.916643311332438e-07,
|
| 2515 |
+
"loss": 1.2566,
|
| 2516 |
+
"step": 1740
|
| 2517 |
+
},
|
| 2518 |
+
{
|
| 2519 |
+
"epoch": 0.8615156751419403,
|
| 2520 |
+
"grad_norm": 0.5844649067792757,
|
| 2521 |
+
"learning_rate": 5.71487723695427e-07,
|
| 2522 |
+
"loss": 1.2176,
|
| 2523 |
+
"step": 1745
|
| 2524 |
+
},
|
| 2525 |
+
{
|
| 2526 |
+
"epoch": 0.8639842014317453,
|
| 2527 |
+
"grad_norm": 0.570757598339604,
|
| 2528 |
+
"learning_rate": 5.516402695259165e-07,
|
| 2529 |
+
"loss": 1.2111,
|
| 2530 |
+
"step": 1750
|
| 2531 |
+
},
|
| 2532 |
+
{
|
| 2533 |
+
"epoch": 0.8664527277215502,
|
| 2534 |
+
"grad_norm": 0.6101964731318252,
|
| 2535 |
+
"learning_rate": 5.321234438050893e-07,
|
| 2536 |
+
"loss": 1.2552,
|
| 2537 |
+
"step": 1755
|
| 2538 |
+
},
|
| 2539 |
+
{
|
| 2540 |
+
"epoch": 0.8689212540113552,
|
| 2541 |
+
"grad_norm": 0.6114031483570134,
|
| 2542 |
+
"learning_rate": 5.12938697139056e-07,
|
| 2543 |
+
"loss": 1.2339,
|
| 2544 |
+
"step": 1760
|
| 2545 |
+
},
|
| 2546 |
+
{
|
| 2547 |
+
"epoch": 0.8713897803011602,
|
| 2548 |
+
"grad_norm": 0.5640524033820485,
|
| 2549 |
+
"learning_rate": 4.940874554518465e-07,
|
| 2550 |
+
"loss": 1.2594,
|
| 2551 |
+
"step": 1765
|
| 2552 |
+
},
|
| 2553 |
+
{
|
| 2554 |
+
"epoch": 0.8738583065909652,
|
| 2555 |
+
"grad_norm": 0.6433079417694005,
|
| 2556 |
+
"learning_rate": 4.755711198794233e-07,
|
| 2557 |
+
"loss": 1.2854,
|
| 2558 |
+
"step": 1770
|
| 2559 |
+
},
|
| 2560 |
+
{
|
| 2561 |
+
"epoch": 0.8763268328807702,
|
| 2562 |
+
"grad_norm": 0.604973387553276,
|
| 2563 |
+
"learning_rate": 4.573910666655429e-07,
|
| 2564 |
+
"loss": 1.3237,
|
| 2565 |
+
"step": 1775
|
| 2566 |
+
},
|
| 2567 |
+
{
|
| 2568 |
+
"epoch": 0.8787953591705752,
|
| 2569 |
+
"grad_norm": 0.5628418770325067,
|
| 2570 |
+
"learning_rate": 4.395486470594645e-07,
|
| 2571 |
+
"loss": 1.1982,
|
| 2572 |
+
"step": 1780
|
| 2573 |
+
},
|
| 2574 |
+
{
|
| 2575 |
+
"epoch": 0.8812638854603801,
|
| 2576 |
+
"grad_norm": 0.6659219563445046,
|
| 2577 |
+
"learning_rate": 4.220451872155179e-07,
|
| 2578 |
+
"loss": 1.2309,
|
| 2579 |
+
"step": 1785
|
| 2580 |
+
},
|
| 2581 |
+
{
|
| 2582 |
+
"epoch": 0.8837324117501851,
|
| 2583 |
+
"grad_norm": 0.5361789546629312,
|
| 2584 |
+
"learning_rate": 4.048819880945337e-07,
|
| 2585 |
+
"loss": 1.199,
|
| 2586 |
+
"step": 1790
|
| 2587 |
+
},
|
| 2588 |
+
{
|
| 2589 |
+
"epoch": 0.8862009380399901,
|
| 2590 |
+
"grad_norm": 0.5558192723511216,
|
| 2591 |
+
"learning_rate": 3.880603253671522e-07,
|
| 2592 |
+
"loss": 1.2263,
|
| 2593 |
+
"step": 1795
|
| 2594 |
+
},
|
| 2595 |
+
{
|
| 2596 |
+
"epoch": 0.8886694643297951,
|
| 2597 |
+
"grad_norm": 0.5634804859248715,
|
| 2598 |
+
"learning_rate": 3.7158144931900395e-07,
|
| 2599 |
+
"loss": 1.2752,
|
| 2600 |
+
"step": 1800
|
| 2601 |
+
},
|
| 2602 |
+
{
|
| 2603 |
+
"epoch": 0.8886694643297951,
|
| 2604 |
+
"eval_loss": 1.2087970972061157,
|
| 2605 |
+
"eval_runtime": 2557.7862,
|
| 2606 |
+
"eval_samples_per_second": 1.564,
|
| 2607 |
+
"eval_steps_per_second": 0.131,
|
| 2608 |
+
"step": 1800
|
| 2609 |
+
},
|
| 2610 |
+
{
|
| 2611 |
+
"epoch": 0.8911379906196001,
|
| 2612 |
+
"grad_norm": 0.6032610406878897,
|
| 2613 |
+
"learning_rate": 3.5544658475778317e-07,
|
| 2614 |
+
"loss": 1.1999,
|
| 2615 |
+
"step": 1805
|
| 2616 |
+
},
|
| 2617 |
+
{
|
| 2618 |
+
"epoch": 0.8936065169094051,
|
| 2619 |
+
"grad_norm": 0.6216254522630721,
|
| 2620 |
+
"learning_rate": 3.396569309222114e-07,
|
| 2621 |
+
"loss": 1.2339,
|
| 2622 |
+
"step": 1810
|
| 2623 |
+
},
|
| 2624 |
+
{
|
| 2625 |
+
"epoch": 0.89607504319921,
|
| 2626 |
+
"grad_norm": 0.5807256981071689,
|
| 2627 |
+
"learning_rate": 3.2421366139290423e-07,
|
| 2628 |
+
"loss": 1.3057,
|
| 2629 |
+
"step": 1815
|
| 2630 |
+
},
|
| 2631 |
+
{
|
| 2632 |
+
"epoch": 0.898543569489015,
|
| 2633 |
+
"grad_norm": 0.5211008570948544,
|
| 2634 |
+
"learning_rate": 3.091179240051462e-07,
|
| 2635 |
+
"loss": 1.2022,
|
| 2636 |
+
"step": 1820
|
| 2637 |
+
},
|
| 2638 |
+
{
|
| 2639 |
+
"epoch": 0.90101209577882,
|
| 2640 |
+
"grad_norm": 0.5525058863296126,
|
| 2641 |
+
"learning_rate": 2.943708407635704e-07,
|
| 2642 |
+
"loss": 1.2048,
|
| 2643 |
+
"step": 1825
|
| 2644 |
+
},
|
| 2645 |
+
{
|
| 2646 |
+
"epoch": 0.903480622068625,
|
| 2647 |
+
"grad_norm": 0.6377145176064325,
|
| 2648 |
+
"learning_rate": 2.799735077587695e-07,
|
| 2649 |
+
"loss": 1.213,
|
| 2650 |
+
"step": 1830
|
| 2651 |
+
},
|
| 2652 |
+
{
|
| 2653 |
+
"epoch": 0.90594914835843,
|
| 2654 |
+
"grad_norm": 0.5813161900855606,
|
| 2655 |
+
"learning_rate": 2.659269950858273e-07,
|
| 2656 |
+
"loss": 1.33,
|
| 2657 |
+
"step": 1835
|
| 2658 |
+
},
|
| 2659 |
+
{
|
| 2660 |
+
"epoch": 0.908417674648235,
|
| 2661 |
+
"grad_norm": 0.6256712692686102,
|
| 2662 |
+
"learning_rate": 2.5223234676478193e-07,
|
| 2663 |
+
"loss": 1.2418,
|
| 2664 |
+
"step": 1840
|
| 2665 |
+
},
|
| 2666 |
+
{
|
| 2667 |
+
"epoch": 0.91088620093804,
|
| 2668 |
+
"grad_norm": 0.598042344925788,
|
| 2669 |
+
"learning_rate": 2.3889058066302873e-07,
|
| 2670 |
+
"loss": 1.2928,
|
| 2671 |
+
"step": 1845
|
| 2672 |
+
},
|
| 2673 |
+
{
|
| 2674 |
+
"epoch": 0.9133547272278449,
|
| 2675 |
+
"grad_norm": 0.6144058961581507,
|
| 2676 |
+
"learning_rate": 2.2590268841966357e-07,
|
| 2677 |
+
"loss": 1.2522,
|
| 2678 |
+
"step": 1850
|
| 2679 |
+
},
|
| 2680 |
+
{
|
| 2681 |
+
"epoch": 0.9158232535176499,
|
| 2682 |
+
"grad_norm": 0.6086868817654493,
|
| 2683 |
+
"learning_rate": 2.132696353717839e-07,
|
| 2684 |
+
"loss": 1.2275,
|
| 2685 |
+
"step": 1855
|
| 2686 |
+
},
|
| 2687 |
+
{
|
| 2688 |
+
"epoch": 0.9182917798074549,
|
| 2689 |
+
"grad_norm": 0.6193803813904503,
|
| 2690 |
+
"learning_rate": 2.0099236048273407e-07,
|
| 2691 |
+
"loss": 1.2102,
|
| 2692 |
+
"step": 1860
|
| 2693 |
+
},
|
| 2694 |
+
{
|
| 2695 |
+
"epoch": 0.9207603060972599,
|
| 2696 |
+
"grad_norm": 0.6206660621687174,
|
| 2697 |
+
"learning_rate": 1.890717762723182e-07,
|
| 2698 |
+
"loss": 1.2413,
|
| 2699 |
+
"step": 1865
|
| 2700 |
+
},
|
| 2701 |
+
{
|
| 2702 |
+
"epoch": 0.9232288323870649,
|
| 2703 |
+
"grad_norm": 0.5195254310690817,
|
| 2704 |
+
"learning_rate": 1.7750876874897627e-07,
|
| 2705 |
+
"loss": 1.2536,
|
| 2706 |
+
"step": 1870
|
| 2707 |
+
},
|
| 2708 |
+
{
|
| 2709 |
+
"epoch": 0.9256973586768699,
|
| 2710 |
+
"grad_norm": 0.6172193600635592,
|
| 2711 |
+
"learning_rate": 1.6630419734393e-07,
|
| 2712 |
+
"loss": 1.1877,
|
| 2713 |
+
"step": 1875
|
| 2714 |
+
},
|
| 2715 |
+
{
|
| 2716 |
+
"epoch": 0.9281658849666748,
|
| 2717 |
+
"grad_norm": 0.5854056073690375,
|
| 2718 |
+
"learning_rate": 1.554588948473068e-07,
|
| 2719 |
+
"loss": 1.2694,
|
| 2720 |
+
"step": 1880
|
| 2721 |
+
},
|
| 2722 |
+
{
|
| 2723 |
+
"epoch": 0.9306344112564798,
|
| 2724 |
+
"grad_norm": 0.5939692455470944,
|
| 2725 |
+
"learning_rate": 1.4497366734623874e-07,
|
| 2726 |
+
"loss": 1.2223,
|
| 2727 |
+
"step": 1885
|
| 2728 |
+
},
|
| 2729 |
+
{
|
| 2730 |
+
"epoch": 0.9331029375462848,
|
| 2731 |
+
"grad_norm": 0.558716522853661,
|
| 2732 |
+
"learning_rate": 1.3484929416495096e-07,
|
| 2733 |
+
"loss": 1.1465,
|
| 2734 |
+
"step": 1890
|
| 2735 |
+
},
|
| 2736 |
+
{
|
| 2737 |
+
"epoch": 0.9355714638360898,
|
| 2738 |
+
"grad_norm": 0.601778856283905,
|
| 2739 |
+
"learning_rate": 1.2508652780683916e-07,
|
| 2740 |
+
"loss": 1.2618,
|
| 2741 |
+
"step": 1895
|
| 2742 |
+
},
|
| 2743 |
+
{
|
| 2744 |
+
"epoch": 0.9380399901258949,
|
| 2745 |
+
"grad_norm": 0.5724230357863298,
|
| 2746 |
+
"learning_rate": 1.1568609389853546e-07,
|
| 2747 |
+
"loss": 1.199,
|
| 2748 |
+
"step": 1900
|
| 2749 |
+
},
|
| 2750 |
+
{
|
| 2751 |
+
"epoch": 0.9405085164156999,
|
| 2752 |
+
"grad_norm": 0.5858685464797397,
|
| 2753 |
+
"learning_rate": 1.0664869113598097e-07,
|
| 2754 |
+
"loss": 1.2416,
|
| 2755 |
+
"step": 1905
|
| 2756 |
+
},
|
| 2757 |
+
{
|
| 2758 |
+
"epoch": 0.9429770427055049,
|
| 2759 |
+
"grad_norm": 0.5955002776535666,
|
| 2760 |
+
"learning_rate": 9.7974991232489e-08,
|
| 2761 |
+
"loss": 1.2621,
|
| 2762 |
+
"step": 1910
|
| 2763 |
+
},
|
| 2764 |
+
{
|
| 2765 |
+
"epoch": 0.9454455689953098,
|
| 2766 |
+
"grad_norm": 0.6031053768787782,
|
| 2767 |
+
"learning_rate": 8.966563886882107e-08,
|
| 2768 |
+
"loss": 1.2966,
|
| 2769 |
+
"step": 1915
|
| 2770 |
+
},
|
| 2771 |
+
{
|
| 2772 |
+
"epoch": 0.9479140952851148,
|
| 2773 |
+
"grad_norm": 0.5626513433181811,
|
| 2774 |
+
"learning_rate": 8.172125164527312e-08,
|
| 2775 |
+
"loss": 1.197,
|
| 2776 |
+
"step": 1920
|
| 2777 |
+
},
|
| 2778 |
+
{
|
| 2779 |
+
"epoch": 0.9503826215749198,
|
| 2780 |
+
"grad_norm": 0.6147790631492948,
|
| 2781 |
+
"learning_rate": 7.414242003576876e-08,
|
| 2782 |
+
"loss": 1.2476,
|
| 2783 |
+
"step": 1925
|
| 2784 |
+
},
|
| 2785 |
+
{
|
| 2786 |
+
"epoch": 0.9528511478647248,
|
| 2787 |
+
"grad_norm": 0.6387128598756113,
|
| 2788 |
+
"learning_rate": 6.692970734397176e-08,
|
| 2789 |
+
"loss": 1.2717,
|
| 2790 |
+
"step": 1930
|
| 2791 |
+
},
|
| 2792 |
+
{
|
| 2793 |
+
"epoch": 0.9553196741545298,
|
| 2794 |
+
"grad_norm": 0.58519229057596,
|
| 2795 |
+
"learning_rate": 6.0083649661421e-08,
|
| 2796 |
+
"loss": 1.2427,
|
| 2797 |
+
"step": 1935
|
| 2798 |
+
},
|
| 2799 |
+
{
|
| 2800 |
+
"epoch": 0.9577882004443348,
|
| 2801 |
+
"grad_norm": 0.5732049204953203,
|
| 2802 |
+
"learning_rate": 5.360475582768088e-08,
|
| 2803 |
+
"loss": 1.2499,
|
| 2804 |
+
"step": 1940
|
| 2805 |
+
},
|
| 2806 |
+
{
|
| 2807 |
+
"epoch": 0.9602567267341398,
|
| 2808 |
+
"grad_norm": 0.5510115335869762,
|
| 2809 |
+
"learning_rate": 4.7493507392524226e-08,
|
| 2810 |
+
"loss": 1.1837,
|
| 2811 |
+
"step": 1945
|
| 2812 |
+
},
|
| 2813 |
+
{
|
| 2814 |
+
"epoch": 0.9627252530239447,
|
| 2815 |
+
"grad_norm": 0.5959129330379044,
|
| 2816 |
+
"learning_rate": 4.175035858013987e-08,
|
| 2817 |
+
"loss": 1.2595,
|
| 2818 |
+
"step": 1950
|
| 2819 |
+
},
|
| 2820 |
+
{
|
| 2821 |
+
"epoch": 0.9651937793137497,
|
| 2822 |
+
"grad_norm": 0.6525575790551825,
|
| 2823 |
+
"learning_rate": 3.637573625537183e-08,
|
| 2824 |
+
"loss": 1.3283,
|
| 2825 |
+
"step": 1955
|
| 2826 |
+
},
|
| 2827 |
+
{
|
| 2828 |
+
"epoch": 0.9676623056035547,
|
| 2829 |
+
"grad_norm": 0.6761446719619785,
|
| 2830 |
+
"learning_rate": 3.13700398919925e-08,
|
| 2831 |
+
"loss": 1.2633,
|
| 2832 |
+
"step": 1960
|
| 2833 |
+
},
|
| 2834 |
+
{
|
| 2835 |
+
"epoch": 0.9701308318933597,
|
| 2836 |
+
"grad_norm": 0.5705669812908541,
|
| 2837 |
+
"learning_rate": 2.673364154301028e-08,
|
| 2838 |
+
"loss": 1.2446,
|
| 2839 |
+
"step": 1965
|
| 2840 |
+
},
|
| 2841 |
+
{
|
| 2842 |
+
"epoch": 0.9725993581831647,
|
| 2843 |
+
"grad_norm": 0.6197155608101478,
|
| 2844 |
+
"learning_rate": 2.2466885813018925e-08,
|
| 2845 |
+
"loss": 1.2492,
|
| 2846 |
+
"step": 1970
|
| 2847 |
+
},
|
| 2848 |
+
{
|
| 2849 |
+
"epoch": 0.9750678844729697,
|
| 2850 |
+
"grad_norm": 0.5667304098455904,
|
| 2851 |
+
"learning_rate": 1.857008983258135e-08,
|
| 2852 |
+
"loss": 1.2485,
|
| 2853 |
+
"step": 1975
|
| 2854 |
+
},
|
| 2855 |
+
{
|
| 2856 |
+
"epoch": 0.9775364107627746,
|
| 2857 |
+
"grad_norm": 0.6113665999543747,
|
| 2858 |
+
"learning_rate": 1.504354323466073e-08,
|
| 2859 |
+
"loss": 1.2573,
|
| 2860 |
+
"step": 1980
|
| 2861 |
+
},
|
| 2862 |
+
{
|
| 2863 |
+
"epoch": 0.9800049370525796,
|
| 2864 |
+
"grad_norm": 0.5726714283406965,
|
| 2865 |
+
"learning_rate": 1.188750813309214e-08,
|
| 2866 |
+
"loss": 1.2264,
|
| 2867 |
+
"step": 1985
|
| 2868 |
+
},
|
| 2869 |
+
{
|
| 2870 |
+
"epoch": 0.9824734633423846,
|
| 2871 |
+
"grad_norm": 0.5521047354644366,
|
| 2872 |
+
"learning_rate": 9.102219103103161e-09,
|
| 2873 |
+
"loss": 1.2194,
|
| 2874 |
+
"step": 1990
|
| 2875 |
+
},
|
| 2876 |
+
{
|
| 2877 |
+
"epoch": 0.9849419896321896,
|
| 2878 |
+
"grad_norm": 0.6819693929722572,
|
| 2879 |
+
"learning_rate": 6.687883163873921e-09,
|
| 2880 |
+
"loss": 1.244,
|
| 2881 |
+
"step": 1995
|
| 2882 |
+
},
|
| 2883 |
+
{
|
| 2884 |
+
"epoch": 0.9874105159219946,
|
| 2885 |
+
"grad_norm": 0.6016814387388122,
|
| 2886 |
+
"learning_rate": 4.644679763155524e-09,
|
| 2887 |
+
"loss": 1.2701,
|
| 2888 |
+
"step": 2000
|
| 2889 |
+
},
|
| 2890 |
+
{
|
| 2891 |
+
"epoch": 0.9874105159219946,
|
| 2892 |
+
"eval_loss": 1.208633542060852,
|
| 2893 |
+
"eval_runtime": 2553.7159,
|
| 2894 |
+
"eval_samples_per_second": 1.566,
|
| 2895 |
+
"eval_steps_per_second": 0.131,
|
| 2896 |
+
"step": 2000
|
| 2897 |
+
},
|
| 2898 |
+
{
|
| 2899 |
+
"epoch": 0.9898790422117996,
|
| 2900 |
+
"grad_norm": 0.5854483828292536,
|
| 2901 |
+
"learning_rate": 2.97276076392905e-09,
|
| 2902 |
+
"loss": 1.2735,
|
| 2903 |
+
"step": 2005
|
| 2904 |
+
},
|
| 2905 |
+
{
|
| 2906 |
+
"epoch": 0.9923475685016045,
|
| 2907 |
+
"grad_norm": 0.6149856349841143,
|
| 2908 |
+
"learning_rate": 1.6722504331195822e-09,
|
| 2909 |
+
"loss": 1.1829,
|
| 2910 |
+
"step": 2010
|
| 2911 |
+
},
|
| 2912 |
+
{
|
| 2913 |
+
"epoch": 0.9948160947914095,
|
| 2914 |
+
"grad_norm": 0.5776580228856067,
|
| 2915 |
+
"learning_rate": 7.432454323597071e-10,
|
| 2916 |
+
"loss": 1.2584,
|
| 2917 |
+
"step": 2015
|
| 2918 |
+
},
|
| 2919 |
+
{
|
| 2920 |
+
"epoch": 0.9972846210812145,
|
| 2921 |
+
"grad_norm": 0.5955477076581019,
|
| 2922 |
+
"learning_rate": 1.8581481080415242e-10,
|
| 2923 |
+
"loss": 1.1737,
|
| 2924 |
+
"step": 2020
|
| 2925 |
+
},
|
| 2926 |
+
{
|
| 2927 |
+
"epoch": 0.9997531473710195,
|
| 2928 |
+
"grad_norm": 0.6070167910291095,
|
| 2929 |
+
"learning_rate": 0.0,
|
| 2930 |
+
"loss": 1.1858,
|
| 2931 |
+
"step": 2025
|
| 2932 |
+
},
|
| 2933 |
+
{
|
| 2934 |
+
"epoch": 0.9997531473710195,
|
| 2935 |
+
"step": 2025,
|
| 2936 |
+
"total_flos": 4526278881050624.0,
|
| 2937 |
+
"train_loss": 1.270192005722611,
|
| 2938 |
+
"train_runtime": 113933.3906,
|
| 2939 |
+
"train_samples_per_second": 0.427,
|
| 2940 |
+
"train_steps_per_second": 0.018
|
| 2941 |
+
}
|
| 2942 |
+
],
|
| 2943 |
+
"logging_steps": 5,
|
| 2944 |
+
"max_steps": 2025,
|
| 2945 |
+
"num_input_tokens_seen": 0,
|
| 2946 |
+
"num_train_epochs": 1,
|
| 2947 |
+
"save_steps": 200,
|
| 2948 |
+
"total_flos": 4526278881050624.0,
|
| 2949 |
+
"train_batch_size": 4,
|
| 2950 |
+
"trial_name": null,
|
| 2951 |
+
"trial_params": null
|
| 2952 |
+
}
|