Update README.md
Browse files
README.md
CHANGED
|
@@ -44,7 +44,6 @@ Script used for training is avaiable [here](https://github.com/harpomaxx/llm-fin
|
|
| 44 |
|
| 45 |
### Training Arguments:
|
| 46 |
|
| 47 |
-
- **Output Directory**: `./results`
|
| 48 |
- **Batch Size**: 4 (per device)
|
| 49 |
- **Gradient Accumulation Steps**: 2
|
| 50 |
- **Number of Epochs**: 10
|
|
@@ -57,6 +56,28 @@ Script used for training is avaiable [here](https://github.com/harpomaxx/llm-fin
|
|
| 57 |
- **Save Steps**: 250
|
| 58 |
- **FP16 Precision**: Enabled
|
| 59 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 60 |
## Usage
|
| 61 |
|
| 62 |
```python
|
|
|
|
| 44 |
|
| 45 |
### Training Arguments:
|
| 46 |
|
|
|
|
| 47 |
- **Batch Size**: 4 (per device)
|
| 48 |
- **Gradient Accumulation Steps**: 2
|
| 49 |
- **Number of Epochs**: 10
|
|
|
|
| 56 |
- **Save Steps**: 250
|
| 57 |
- **FP16 Precision**: Enabled
|
| 58 |
|
| 59 |
+
|
| 60 |
+
### Training information from wandb
|
| 61 |
+
|
| 62 |
+
- **train/total_flos:** 72,761,086,854,758,400
|
| 63 |
+
- **train/train_loss:** 1.5557164267259171
|
| 64 |
+
- **train/train_runtime:** 5892.7285 seconds
|
| 65 |
+
- **train/train_steps_per_second:** 4.248
|
| 66 |
+
- **_runtime:** 5891.33976650238 seconds
|
| 67 |
+
- **_timestamp:** 1,695,390,058.0198596
|
| 68 |
+
- **train/epoch:** 10
|
| 69 |
+
- **train/global_step:** 25,030
|
| 70 |
+
- **train/learning_rate:** 8.371592860045851e-12
|
| 71 |
+
- **train/train_samples_per_second:** 33.977
|
| 72 |
+
- **_step:** 2,503
|
| 73 |
+
- **_wandb.runtime:** 5890 seconds
|
| 74 |
+
- **train/loss:** 1.4114
|
| 75 |
+
|
| 76 |
+
|
| 77 |
+
|
| 78 |
+
|
| 79 |
+
|
| 80 |
+
|
| 81 |
## Usage
|
| 82 |
|
| 83 |
```python
|