Commit
•
6b79b2a
1
Parent(s):
3726114
add finetuning examples
Browse files
README.md
CHANGED
|
@@ -93,6 +93,8 @@ for i, t in enumerate(generated_text):
|
|
| 93 |
|
| 94 |
To quickly test your software without waiting for the huge model to download/load you can use `HuggingFaceM4/tiny-random-idefics` - it hasn't been trained and has random weights but it is very useful for quick testing.
|
| 95 |
|
|
|
|
|
|
|
| 96 |
# Training Details
|
| 97 |
|
| 98 |
We closely follow the training procedure layed out in [Flamingo](https://huggingface.co/papers/2204.14198). We combine two open-source pre-trained models ([laion/CLIP-ViT-H-14-laion2B-s32B-b79K](https://huggingface.co/laion/CLIP-ViT-H-14-laion2B-s32B-b79K) and [huggyllama/llama-65b](https://huggingface.co/huggyllama/llama-65b)) by initializing new Transformer blocks. The pre-trained backbones are frozen while we train the newly initialized parameters.
|
|
|
|
| 93 |
|
| 94 |
To quickly test your software without waiting for the huge model to download/load you can use `HuggingFaceM4/tiny-random-idefics` - it hasn't been trained and has random weights but it is very useful for quick testing.
|
| 95 |
|
| 96 |
+
This [tutorial](https://github.com/huggingface/notebooks/pull/418/) shows a simple example to fine-tune IDEFICS on custom data. This [colab notebook](TODO) showcases how to do the fine-tuning in 4bits precision. TODO: change to the correct link once it's merged.
|
| 97 |
+
|
| 98 |
# Training Details
|
| 99 |
|
| 100 |
We closely follow the training procedure layed out in [Flamingo](https://huggingface.co/papers/2204.14198). We combine two open-source pre-trained models ([laion/CLIP-ViT-H-14-laion2B-s32B-b79K](https://huggingface.co/laion/CLIP-ViT-H-14-laion2B-s32B-b79K) and [huggyllama/llama-65b](https://huggingface.co/huggyllama/llama-65b)) by initializing new Transformer blocks. The pre-trained backbones are frozen while we train the newly initialized parameters.
|