Update README.md

README.md

@@ -17,8 +17,18 @@ This repo is the result of converting to GGML and quantising.
 
 ## Repositories available
 
-* [MPT-7B: 4-bit, 5-bit and 8-bit GGML models for CPU (+CUDA) inference](https://huggingface.co/TheBloke/MPT-7B).
-* [MPT-7B-Instruct: 4-bit, 5-bit and 8-bit GGML models for CPU (+CUDA) inference](https://huggingface.co/TheBloke/MPT-7B-Instruct).
+* [MPT-7B: 4-bit, 5-bit and 8-bit GGML models for CPU (+CUDA) inference](https://huggingface.co/TheBloke/MPT-7B-GGML).
+* [MPT-7B-Instruct: 4-bit, 5-bit and 8-bit GGML models for CPU (+CUDA) inference](https://huggingface.co/TheBloke/MPT-7B-Instruct-GGML).
+
+## Provided files
+| Name | Quant method | Bits | Size | RAM required | Use case |
+| ---- | ---- | ---- | ---- | ---- | ----- |
+`mpt7b-instructggmlv2.ggmlv2.q4_0.bin` | q4_0 | 4bit | 4.21GB | 7.0GB | 4-bit. |
+`mpt7b-instructggmlv2.ggmlv2.q4_1.bin` | q4_0 | 4bit | 4.63GB | 7.5GB | 4-bit. Higher accuracy than q4_0 but not as high as q5_0. However has quicker inference than q5 models. |
+`mpt7b-instructggmlv2.ggmlv2.q5_0.bin` | q5_0 | 5bit | 4.63GB | 7.5GB | 5-bit. Higher accuracy, higher resource usage and slower inference. |
+`mpt7b-instructggmlv2.ggmlv2.q5_1.bin` | q5_1 | 5bit | 5.06GB | 7.5GB | 5-bit. Even higher accuracy, and higher resource usage and slower inference. |
+`mpt7b-instructggmlv2.ggmlv2.q8_0.bin` | q8_0 | 8bit | 7.58GB | 9.0GB | 8-bit. Almost indistinguishable from float16. Huge resource use and slow. Not recommended for normal use. |
+`mpt7b-instructggmlv2.ggmlv2.fp16.bin` | fp16 | 16bit | GB | GB | Full 16-bit. |
 
 ## Compatibilty
 
@@ -44,16 +54,6 @@ bin/mpt -m /path/to/mpt7b-instructggmlv2.ggmlv2.q4_0.bin -t 8 -n 512 -p "Write a
 
 Please see the ggml repo for other build options.
 
-## Provided files
-| Name | Quant method | Bits | Size | RAM required | Use case |
-| ---- | ---- | ---- | ---- | ---- | ----- |
-`mpt7b-instructggmlv2.ggmlv2.q4_0.bin` | q4_0 | 4bit | 4.21GB | 7.0GB | 4-bit. |
-`mpt7b-instructggmlv2.ggmlv2.q4_1.bin` | q4_0 | 4bit | 4.63GB | 7.5GB | 4-bit. Higher accuracy than q4_0 but not as high as q5_0. However has quicker inference than q5 models. |
-`mpt7b-instructggmlv2.ggmlv2.q5_0.bin` | q5_0 | 5bit | 4.63GB | 7.5GB | 5-bit. Higher accuracy, higher resource usage and slower inference. |
-`mpt7b-instructggmlv2.ggmlv2.q5_1.bin` | q5_1 | 5bit | 5.06GB | 7.5GB | 5-bit. Even higher accuracy, and higher resource usage and slower inference. |
-`mpt7b-instructggmlv2.ggmlv2.q8_0.bin` | q8_0 | 8bit | 7.58GB | 9.0GB | 8-bit. Almost indistinguishable from float16. Huge resource use and slow. Not recommended for normal use. |
-`mpt7b-instructggmlv2.ggmlv2.fp16.bin` | fp16 | 16bit | GB | GB | Full 16-bit. |
-
 
 # Original model card: MPT-7B-Instruct