From 85522184917cd701863147aebcc5619efbf9bf06 Mon Sep 17 00:00:00 2001
From: NanoCode012 <kevinvong@rocketmail.com>
Date: Thu, 25 May 2023 17:04:44 +0900
Subject: [PATCH] Improve Inference instruction

---
 README.md | 14 +++++++++-----
 1 file changed, 9 insertions(+), 5 deletions(-)

diff --git a/README.md b/README.md
index 969708d47..531270dcc 100644
--- a/README.md
+++ b/README.md
@@ -317,12 +317,16 @@ accelerate launch scripts/finetune.py configs/your_config.yml
 
 ### Inference
 
-Add `--inference` flag to train command above
+Pass the appropriate flag to the train command:
 
-If you are inferencing a pretrained LORA, pass 
-```bash
---lora_model_dir ./completed-model
-```
+- Pretrained LORA:
+  ```bash
+  --inference --lora_model_dir ./completed-model
+  ```
+- Full weights finetune:
+  ```bash
+  --inference --base_model ./completed-model
+  ```
 
 ### Merge LORA to base