Update README.md
Browse files
README.md
CHANGED
|
@@ -14,7 +14,43 @@ The following `bitsandbytes` quantization config was used during training:
|
|
| 14 |
- bnb_4bit_quant_type: fp4
|
| 15 |
- bnb_4bit_use_double_quant: False
|
| 16 |
- bnb_4bit_compute_dtype: float32
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 17 |
### Framework versions
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 18 |
|
|
|
|
| 19 |
|
| 20 |
-
|
|
|
|
| 14 |
- bnb_4bit_quant_type: fp4
|
| 15 |
- bnb_4bit_use_double_quant: False
|
| 16 |
- bnb_4bit_compute_dtype: float32
|
| 17 |
+
|
| 18 |
+
### Model Description
|
| 19 |
+
|
| 20 |
+
For more information on how it was created, check out the following link: https://github.com/DunnBC22/NLP_Projects/blob/main/OPT%20Models/Grade%20School%20Math%20Instructions%20Fine-Tune%20OPT.ipynb.
|
| 21 |
+
|
| 22 |
+
### Intended uses & limitations
|
| 23 |
+
|
| 24 |
+
This is intended to show the possibilities. It is mainly limited by the input data.
|
| 25 |
+
|
| 26 |
+
### Training & Evaluation Dataset
|
| 27 |
+
|
| 28 |
+
Dataset Source: https://huggingface.co/datasets/qwedsacf/grade-school-math-instructions
|
| 29 |
+
|
| 30 |
+
### Hyperparameters Used
|
| 31 |
+
|
| 32 |
+
| Hyperperameter | Value |
|
| 33 |
+
|:-----:|:-----:|
|
| 34 |
+
| Model Checkpoint | facebook/opt-2.7b |
|
| 35 |
+
| per_device_train_batch_size | 4 |
|
| 36 |
+
| gradient_accumulation_steps | 4 |
|
| 37 |
+
| fp16 | True |
|
| 38 |
+
| warmup_steps | 225 |
|
| 39 |
+
| learning_rate | 2e-4 |
|
| 40 |
+
| Training Steps | 450
|
| 41 |
+
|
| 42 |
+
|
| 43 |
### Framework versions
|
| 44 |
+
|
| 45 |
+
| Library | Version |
|
| 46 |
+
|:-----:|:-----:|
|
| 47 |
+
| Python | 3.10.1 |
|
| 48 |
+
| Torch | 2.0.1+cu118 |
|
| 49 |
+
| Datasets | 2.14.4 |
|
| 50 |
+
| Transformer | 4.31.0 |
|
| 51 |
+
| PEFT | 0.4.0
|
| 52 |
+
|
| 53 |
|
| 54 |
+
### Metric
|
| 55 |
|
| 56 |
+
Perplexity = 6.35
|