Update README.md

README.md

@@ -16,7 +16,15 @@ These are the checkpoints for the first ever QLoRA for ESM-2! They haven't been
 You can load and use them similarly to the LoRA models. This is the smallest `esm2_t6_8M_UR50D` model, so the metrics aren't great. 
 Scaling to larger models for better metrics is in progress. These checkpoints were trained using 
 [the 600K dataset](https://huggingface.co/datasets/AmelieSchreiber/600K_data). To replicate the training of QLoRA for ESM-2 models, 
-you can use the `conda-environment.yml` file. 
+you can use the `conda-environment.yml` file. However, for the next week or two (28/09/2023) you will need to uninstall transformers
+and use this instead:
+
+```
+pip install --upgrade git+https://github.com/huggingface/transformers.git
+```
+
+Once the transformers library is updated, you should be able to simply use the latest version of transformers and gradient checkpointing 
+will be fully enabled, and QLoRA compatibility should be fully integrated into ESM-2 models. 
 
 ## QLoRA Info