Initial GGML model commit

README.md

@@ -39,6 +39,8 @@ This repo contains GGML format model files for [Jon Durbin's Airoboros L2 70B](h
 
 The GGML format has now been superseded by GGUF. As of August 21st 2023, [llama.cpp](https://github.com/ggerganov/llama.cpp) no longer supports GGML models. Third party clients and libraries are expected to still support it for a time, but many may also drop support.
 
+Please use the GGUF models instead.
+
 ### About GGML
 
 GPU acceleration is now available for Llama 2 70B GGML files, with both CUDA (NVidia) and Metal (macOS). The following clients/libraries are known to work with these files, including with GPU acceleration:
@@ -111,39 +113,8 @@ Refer to the Provided Files table below to see what files use which methods, and
 | [airoboros-l2-70b-2.1.ggmlv3.Q5_0.bin](https://huggingface.co/TheBloke/Airoboros-L2-70B-2.1-GGML/blob/main/airoboros-l2-70b-2.1.ggmlv3.Q5_0.bin) | Q5_0 | 5 | 47.46 GB| 49.96 GB | Original quant method, 5-bit. Higher accuracy, higher resource usage and slower inference. |
 | [airoboros-l2-70b-2.1.ggmlv3.Q5_K_S.bin](https://huggingface.co/TheBloke/Airoboros-L2-70B-2.1-GGML/blob/main/airoboros-l2-70b-2.1.ggmlv3.Q5_K_S.bin) | Q5_K_S | 5 | 47.46 GB| 49.96 GB | New k-quant method. Uses GGML_TYPE_Q5_K for all tensors |
 | [airoboros-l2-70b-2.1.ggmlv3.Q5_K_M.bin](https://huggingface.co/TheBloke/Airoboros-L2-70B-2.1-GGML/blob/main/airoboros-l2-70b-2.1.ggmlv3.Q5_K_M.bin) | Q5_K_M | 5 | 48.75 GB| 51.25 GB | New k-quant method. Uses GGML_TYPE_Q6_K for half of the attention.wv and feed_forward.w2 tensors, else GGML_TYPE_Q5_K |
-| airoboros-l2-70b-2.1.ggmlv3.q5_1.bin | q5_1 | 5 | 51.76 GB | 54.26 GB | Original quant method, 5-bit. Higher accuracy, slower inference than q5_0. |
-| airoboros-l2-70b-2.1.ggmlv3.q6_K.bin | q6_K | 6 | 56.59 GB | 59.09 GB | New k-quant method. Uses GGML_TYPE_Q8_K - 6-bit quantization - for all tensors |
-| airoboros-l2-70b-2.1.ggmlv3.q8_0.bin | q8_0 | 8 | 73.23 GB | 75.73 GB | Original llama.cpp quant method, 8-bit. Almost indistinguishable from float16. High resource use and slow. Not recommended for most users. |
-
-### q5_1, q6_K and q8_0 files require expansion from archive
-
-**Note:** HF does not support uploading files larger than 50GB. Therefore I have uploaded the q6_K and q8_0 files as multi-part ZIP files. They are not compressed, they are just for storing a .bin file in two parts.
 
-<details>
-  <summary>Click for instructions regarding q5_1, q6_K and q8_0 files</summary>
-  
-### q5_1
-Please download:
-* `airoboros-l2-70b-2.1.ggmlv3.q5_1.zip`
-* `airoboros-l2-70b-2.1.ggmlv3.q5_1.z01`
- 
-### q6_K 
-Please download:
-* `airoboros-l2-70b-2.1.ggmlv3.q6_K.zip`
-* `airoboros-l2-70b-2.1.ggmlv3.q6_K.z01`
-
-### q8_0
-Please download:
-* `airoboros-l2-70b-2.1.ggmlv3.q8_0.zip`
-* `airoboros-l2-70b-2.1.ggmlv3.q8_0.z01`
-
-Then extract the .zip archive. This will will expand both parts automatically. On Linux I found I had to use `7zip` - the basic `unzip` tool did not work. Example:
-```
-sudo apt update -y && sudo apt install 7zip
-7zz x airoboros-l2-70b-2.1.ggmlv3.q6_K.zip
-```
-
-</details>
+**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`
 
@@ -209,7 +180,9 @@ And thank you again to a16z for their generous grant.
 
 ### Overview
 
-__*I haven't tested this at all yet, quality could be great or absolute trash, I really don't know, but feel free to try.*__
+__*NOTE: The weights have been re-uploaded as of 2023-08-28 06:57PM EST*__
+
+__*I re-merged the adapter weights (info here: https://twitter.com/jon_durbin/status/1696243076178571474)*__
 
 This is an instruction fine-tuned llama-2 model, using synthetic data generated by [airoboros](https://github.com/jondurbin/airoboros)