New GGMLv3 format for breaking llama.cpp change May 19th commit 2d5db48

README.md

@@ -19,29 +19,28 @@ These files are for CPU (+ CUDA) inference using [llama.cpp](https://github.com/
 * [4bit's GPTQ 4-bit model for GPU inference](https://huggingface.co/4bit/WizardLM-13B-Uncensored-4bit-128g).
 * [Eric's original float16 HF format model for GPU inference and further conversions](https://huggingface.co/ehartford/WizardLM-13B-Uncensored).
 
-## THESE FILES REQUIRE LATEST LLAMA.CPP (May 12th 2023 - commit b9fd7ee)!
+## THE FILES IN MAIN BRANCH REQUIRES LATEST LLAMA.CPP (May 19th 2023 - commit 2d5db48)!
 
-llama.cpp recently made a breaking change to its quantisation methods.
+llama.cpp recently made another breaking change to its quantisation methods - https://github.com/ggerganov/llama.cpp/pull/1508
 
-I have quantised the GGML files in this repo with the latest version. Therefore you will require llama.cpp compiled on May 12th or later (commit `b9fd7ee` or later) to use them.
-
-If you are currently unable to update llama.cpp, eg because you use a UI which hasn't updated yet, you can find a [q5_1 GGML for the older llama.cpp code here](https://huggingface.co/TehVenom/WizardLM-13B-Uncensored-Q5_1-GGML).
+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.
 
+For files compatible with the previous version of llama.cpp, please see branch `previous_llama_ggmlv2`.
 ## Provided files
 | Name | Quant method | Bits | Size | RAM required | Use case |
 | ---- | ---- | ---- | ---- | ---- | ----- |
-`wizardLM-13B-Uncensored.ggml.q4_0.bin` | q4_0 | 4bit | 8.14GB | 10.5GB | 4-bit. |
-`wizardLM-13B-Uncensored.ggml.q4_1.bin` | q4_1 | 5bit | 9.76GB | 12.25GB | 4-bit. Higher accuracy than q4_0 but not as high as q5_0. However has quicker inference than q5 models. |
-`wizardLM-13B-Uncensored.ggml.q5_0.bin` | q5_0 | 5bit | 8.95GB | 11.0GB | 5-bit. Higher accuracy, higher resource usage and slower inference. |
-`wizardLM-13B-Uncensored.ggml.q5_1.bin` | q5_1 | 5bit | 9.76GB | 12.25GB | 5-bit. Even higher accuracy, and higher resource usage and slower inference. |
-`wizardLM-13B-Uncensored.ggml.q8_0.bin` | q8_0 | 8bit | 14.6GB | 17GB | 8-bit. Almost indistinguishable from float16. Huge resource use and slow. Not recommended for normal use. |
+`wizardLM-13B-Uncensored.ggmlv3.q4_0.bin` | q4_0 | 4bit | 8.14GB | 10.5GB | 4-bit. |
+`wizardLM-13B-Uncensored.ggmlv3.q4_1.bin` | q4_1 | 5bit | 9.76GB | 12.25GB | 4-bit. Higher accuracy than q4_0 but not as high as q5_0. However has quicker inference than q5 models. |
+`wizardLM-13B-Uncensored.ggmlv3.q5_0.bin` | q5_0 | 5bit | 8.95GB | 11.0GB | 5-bit. Higher accuracy, higher resource usage and slower inference. |
+`wizardLM-13B-Uncensored.ggmlv3.q5_1.bin` | q5_1 | 5bit | 9.76GB | 12.25GB | 5-bit. Even higher accuracy, and higher resource usage and slower inference. |
+`wizardLM-13B-Uncensored.ggmlv3.q8_0.bin` | q8_0 | 8bit | 14.6GB | 17GB | 8-bit. Almost indistinguishable from float16. Huge resource use and slow. Not recommended for normal use. |
 
 ## How to run in `llama.cpp`
 
 I use the following command line; adjust for your tastes and needs:
 
 ```
-./main -t 8 -m wizardLM-13B-Uncensored.ggml.q5_0.bin --color -c 2048 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "### Instruction: write a story about llamas ### Response:"
+./main -t 8 -m wizardLM-13B-Uncensored.ggmlv3.q5_0.bin --color -c 2048 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "### Instruction: write a story about llamas ### Response:"
 ```
 
 Change `-t 8` to the number of physical CPU cores you have.
@@ -52,6 +51,8 @@ GGML models can be loaded into text-generation-webui by installing the llama.cpp
 
 Further instructions here: [text-generation-webui/docs/llama.cpp-models.md](https://github.com/oobabooga/text-generation-webui/blob/main/docs/llama.cpp-models.md).
 
+Note: at this time text-generation-webui may not support the new May 19th llama.cpp quantisation methods for q4_0, q4_1 and q8_0 files.
+
 # Original model card
 
 This is WizardLM trained with a subset of the dataset - responses that contained alignment / moralizing were removed.  The intent is to train a WizardLM that doesn't have alignment built-in, so that alignment (of any sort) can be added separately with for example with a RLHF LoRA.