Update README.md
Browse files
README.md
CHANGED
|
@@ -85,6 +85,12 @@ special_tokens:
|
|
| 85 |
unk_token: "<unk>"
|
| 86 |
```
|
| 87 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 88 |
## Training procedure
|
| 89 |
|
| 90 |
The following `bitsandbytes` quantization config was used during training:
|
|
|
|
| 85 |
unk_token: "<unk>"
|
| 86 |
```
|
| 87 |
|
| 88 |
+
And then merged with Axolotl via:
|
| 89 |
+
|
| 90 |
+
```
|
| 91 |
+
accelerate launch scripts/finetune.py configs/your_config.yml --merge_lora --lora_model_dir="./completed-model" --load_in_8bit=False --load_in_4bit=False
|
| 92 |
+
```
|
| 93 |
+
|
| 94 |
## Training procedure
|
| 95 |
|
| 96 |
The following `bitsandbytes` quantization config was used during training:
|