diff --git a/Home.md b/Home.md
deleted file mode 100644
index 8448d13..0000000
--- a/Home.md
+++ /dev/null
@@ -1 +0,0 @@
-Welcome to the text-generation-webui wiki!
diff --git a/Using-LoRAs.md b/Using-LoRAs.md
new file mode 100644
index 0000000..91fdaaf
--- /dev/null
+++ b/Using-LoRAs.md
@@ -0,0 +1,54 @@
+Based on https://github.com/tloen/alpaca-lora/
+
+Probably not the best way to do it. WIP. Suggestions are welcome.
+
+## Instructions
+1. Re-install the requirements
+
+```
+pip install -r requirements.txt
+```
+
+2. Download the LoRA
+
+```
+python download-model.py tloen/alpaca-lora-7b
+```
+
+3. Load llama-7b in 8-bit mode (it only seems to work in 8-bit mode, don't ask me why)
+
+```
+python server.py --model llama-7b --load-in-8bit
+```
+
+4. Select the LoRA in the Parameters tab.
+
+## Prompt
+For this particular LoRA, apparently the prompt must be formatted like this:
+
+```
+Below is an instruction that describes a task. Write a response that appropriately completes the request.
+### Instruction:
+Write a Python script that generates text using the transformers library.
+### Response:
+```
+
+Sample output:
+
+```
+Below is an instruction that describes a task. Write a response that appropriately completes the request.
+### Instruction:
+Write a Python script that generates text using the transformers library.
+### Response:
+
+import transformers
+from transformers import AutoTokenizer, AutoModelForCausalLM
+tokenizer = AutoTokenizer.from_pretrained("bert-base-uncased")
+model = AutoModelForCausalLM.from_pretrained("bert-base-uncased")
+texts = ["Hello world", "How are you"]
+for sentence in texts:
+sentence = tokenizer(sentence)
+print(f"Generated {len(sentence)} tokens from '{sentence}'")
+output = model(sentences=sentence).predict()
+print(f"Predicted {len(output)} tokens for '{sentence}':\n{output}")
+```
\ No newline at end of file