Update README.md

README.md

@@ -48,6 +48,21 @@ GGML files are for CPU + GPU inference using [llama.cpp](https://github.com/gger
 * [4-bit, 5-bit, and 8-bit GGML models for CPU+GPU inference](https://huggingface.co/TheBloke/hippogriff-30b-chat-GGML)
 * [Unquantised fp16 model in HF format](https://huggingface.co/openaccess-ai-collective/hippogriff-30b-chat)
 
+## Prompt template
+
+```
+You are a helpful assistant
+USER: prompt goes here
+ASSISTANT:
+```
+or
+
+```
+<|system|> You are a helpful assistant
+<|user|> prompt goes here
+<|model|>
+```
+
 ## THE FILES IN MAIN BRANCH REQUIRES LATEST LLAMA.CPP (May 19th 2023 - commit 2d5db48)!
 
 llama.cpp recently made another breaking change to its quantisation methods - https://github.com/ggerganov/llama.cpp/pull/1508
@@ -55,7 +70,7 @@ llama.cpp recently made another breaking change to its quantisation methods - ht
 I have quantised the GGML files in this repo with the latest version. Therefore you will require llama.cpp compiled on May 19th or later (commit `2d5db48` or later) to use them.
 
 ## Provided files
-| Name | Quant method | Bits | Size | RAM required | Use case |
+| Name | Quant method | Bits | Size | Max RAM required | Use case |
 | ---- | ---- | ---- | ---- | ---- | ----- |
 | hippogriff-30b.ggmlv3.q4_0.bin | q4_0 | 4 | 18.30 GB | 20.80 GB | 4-bit. |
 | hippogriff-30b.ggmlv3.q4_1.bin | q4_1 | 4 | 20.33 GB | 22.83 GB | 4-bit. Higher accuracy than q4_0 but not as high as q5_0. However has quicker inference than q5 models. |
@@ -63,13 +78,14 @@ I have quantised the GGML files in this repo with the latest version. Therefore
 | hippogriff-30b.ggmlv3.q5_1.bin | q5_1 | 5 | 24.40 GB | 26.90 GB | 5-bit. Even higher accuracy, resource usage and slower inference. |
 | hippogriff-30b.ggmlv3.q8_0.bin | q8_0 | 8 | 34.56 GB | 37.06 GB | 8-bit. Almost indistinguishable from float16. Huge resource use and slow. Not recommended for normal use. |
 
+**Note**: the above RAM figures assume no GPU offloading. If layers are offloaded to the GPU, this will reduce RAM usage and use VRAM instead.
 
 ## How to run in `llama.cpp`
 
-I use the following command line; adjust for your tastes and needs:
+Here is an example command line. Adjust for your tastes and needs:
 
 ```
-./main -t 10 -ngl 32 -m hippogriff-30b.ggmlv3.q5_0.bin --color -c 2048 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "### Instruction: Write a story about llamas\n### Response:"
+./main -t 10 -ngl 32 -m hippogriff-30b.ggmlv3.q5_0.bin --color -c 2048 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "<|system|> You are a story writing assistant\n<|user|> prompt\n<|model|>"
 ```
 Change `-t 10` to the number of physical CPU cores you have. For example if your system has 8 cores/16 threads, use `-t 8`.