Upload new k-quant GGML quantised models.

README.md

@@ -1,9 +1,8 @@
 ---
+inference: false
 license: other
-datasets:
-- c-s-ale/alpaca-gpt4-data
-pipeline_tag: text2text-generation
 ---
+
 <!-- header start -->
 <div style="width: 100%;">
     <img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
@@ -18,37 +17,86 @@ pipeline_tag: text2text-generation
 </div>
 <!-- header end -->
 
-## GPT4-Alpaca-LoRA_MLP-65B GGML
+# Chtan's GPT4 Alpaca LoRA MLP 65B GGML
 
-These files are the result of merging the [LoRA weights of chtan's gpt4-alpaca-lora_mlp-65B](https://huggingface.co/chtan/gpt4-alpaca-lora_mlp-65b) with the original Llama 65B model.
+These files are GGML format model files for [Chtan's GPT4 Alpaca LoRA MLP 65B](https://huggingface.co/chtan/gpt4-alpaca-lora_mlp-65b).
 
-This repo contains 4bit and 5bit quantised GGML files for CPU inference using [llama.cpp](https://github.com/ggerganov/llama.cpp).
+GGML files are for CPU + GPU inference using [llama.cpp](https://github.com/ggerganov/llama.cpp) and libraries and UIs which support this format, such as:
+* [text-generation-webui](https://github.com/oobabooga/text-generation-webui)
+* [KoboldCpp](https://github.com/LostRuins/koboldcpp)
+* [ParisNeo/GPT4All-UI](https://github.com/ParisNeo/gpt4all-ui)
+* [llama-cpp-python](https://github.com/abetlen/llama-cpp-python)
+* [ctransformers](https://github.com/marella/ctransformers)
 
 ## Repositories available
 
-* [4bit GPTQ models for GPU inference](https://huggingface.co/TheBloke/gpt4-alpaca-lora_mlp-65B-GPTQ)
-* [4bit and 5bit GGML models for CPU inference in llama.cpp](https://huggingface.co/TheBloke/gpt4-alpaca-lora_mlp-65B-GGML)
-* [float16 unquantised model for GPU inference and further conversions](https://huggingface.co/TheBloke/gpt4-alpaca-lora_mlp-65B-HF)
+* [4-bit GPTQ models for GPU inference](https://huggingface.co/TheBloke/gpt4-alpaca-lora_mlp-65B-GPTQ)
+* [2, 3, 4, 5, 6 and 8-bit GGML models for CPU+GPU inference](https://huggingface.co/TheBloke/gpt4-alpaca-lora_mlp-65B-GGML)
+* [Unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/TheBloke/gpt4-alpaca-lora_mlp-65B-HF)
+
+<!-- compatibility_ggml start -->
+## Compatibility
+
+### Original llama.cpp quant methods: `q4_0, q4_1, q5_0, q5_1, q8_0`
+
+I have quantized these 'original' quantisation methods using an older version of llama.cpp so that they remain compatible with llama.cpp as of May 19th, commit `2d5db48`.
+
+They should be compatible with all current UIs and libraries that use llama.cpp, such as those listed at the top of this README.
+
+### New k-quant methods: `q2_K, q3_K_S, q3_K_M, q3_K_L, q4_K_S, q4_K_M, q5_K_S, q6_K`
 
-## THE FILES IN MAIN BRANCH REQUIRES LATEST LLAMA.CPP (May 19th 2023 - commit 2d5db48)!
+These new quantisation methods are only compatible with llama.cpp as of June 6th, commit `2d43387`.
 
-llama.cpp recently made another breaking change to its quantisation methods - https://github.com/ggerganov/llama.cpp/pull/1508
+They will NOT be compatible with koboldcpp, text-generation-ui, and other UIs and libraries yet. Support is expected to come over the next few days.
 
-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.
+## Explanation of the new k-quant methods
 
-For files compatible with the previous version of llama.cpp, please see branch `previous_llama_ggmlv2`.
+The new methods available are:
+* GGML_TYPE_Q2_K - "type-1" 2-bit quantization in super-blocks containing 16 blocks, each block having 16 weight. Block scales and mins are quantized with 4 bits. This ends up effectively using 2.5625 bits per weight (bpw)
+* GGML_TYPE_Q3_K - "type-0" 3-bit quantization in super-blocks containing 16 blocks, each block having 16 weights. Scales are quantized with 6 bits. This end up using 3.4375 bpw.
+* GGML_TYPE_Q4_K - "type-1" 4-bit quantization in super-blocks containing 8 blocks, each block having 32 weights. Scales and mins are quantized with 6 bits. This ends up using 4.5 bpw.
+* GGML_TYPE_Q5_K - "type-1" 5-bit quantization. Same super-block structure as GGML_TYPE_Q4_K resulting in 5.5 bpw
+* GGML_TYPE_Q6_K - "type-0" 6-bit quantization. Super-blocks with 16 blocks, each block having 16 weights. Scales are quantized with 8 bits. This ends up using 6.5625 bpw
+* GGML_TYPE_Q8_K - "type-0" 8-bit quantization. Only used for quantizing intermediate results. The difference to the existing Q8_0 is that the block size is 256. All 2-6 bit dot products are implemented for this quantization type.
+
+Refer to the Provided Files table below to see what files use which methods, and how.
+<!-- compatibility_ggml end -->
 
 ## Provided files
-| Name | Quant method | Bits | Size | RAM required | Use case |
+| Name | Quant method | Bits | Size | Max RAM required | Use case |
 | ---- | ---- | ---- | ---- | ---- | ----- |
-`gpt4-alpaca-lora_mlp-65B.ggmlv3.q4_0.bin` | q4_0 | 4bit | 40.8GB | 43GB | 4-bit. |
-`gpt4-alpaca-lora_mlp-65B.ggmlv3.q4_1.bin` | q4_0 | 4bit | 44.9GB | 47GB | 4-bit. Higher accuracy than q4_0 but not as high as q5_0. However has quicker inference than q5 models. |
-`gpt4-alpaca-lora_mlp-65B.ggmlv3.q5_0.bin` | q5_0 | 5bit | 44.9GB | 47GB | 5-bit. Higher accuracy, higher resource usage and slower inference.  |
-`gpt4-alpaca-lora_mlp-65B.ggmlv3.q5_1.bin` | q5_1 | 5bit | 49.0GB | 51GB | 5-bit. Even higher accuracy, higher resource usage and slower inference. |
+| gpt4-alpaca-lora_mlp-65B.ggmlv3.q2_K.bin | q2_K | 2 | 27.33 GB | 29.83 GB | New k-quant method. Uses GGML_TYPE_Q4_K for the attention.vw and feed_forward.w2 tensors, GGML_TYPE_Q2_K for the other tensors. |
+| gpt4-alpaca-lora_mlp-65B.ggmlv3.q3_K_L.bin | q3_K_L | 3 | 34.55 GB | 37.05 GB | New k-quant method. Uses GGML_TYPE_Q5_K for the attention.wv, attention.wo, and feed_forward.w2 tensors, else GGML_TYPE_Q3_K |
+| gpt4-alpaca-lora_mlp-65B.ggmlv3.q3_K_M.bin | q3_K_M | 3 | 31.40 GB | 33.90 GB | New k-quant method. Uses GGML_TYPE_Q4_K for the attention.wv, attention.wo, and feed_forward.w2 tensors, else GGML_TYPE_Q3_K |
+| gpt4-alpaca-lora_mlp-65B.ggmlv3.q3_K_S.bin | q3_K_S | 3 | 28.06 GB | 30.56 GB | New k-quant method. Uses GGML_TYPE_Q3_K for all tensors |
+| gpt4-alpaca-lora_mlp-65B.ggmlv3.q4_0.bin | q4_0 | 4 | 36.73 GB | 39.23 GB | Original llama.cpp quant method, 4-bit. |
+| gpt4-alpaca-lora_mlp-65B.ggmlv3.q4_1.bin | q4_1 | 4 | 40.81 GB | 43.31 GB | Original llama.cpp quant method, 4-bit. Higher accuracy than q4_0 but not as high as q5_0. However has quicker inference than q5 models. |
+| gpt4-alpaca-lora_mlp-65B.ggmlv3.q4_K_M.bin | q4_K_M | 4 | 39.28 GB | 41.78 GB | New k-quant method. Uses GGML_TYPE_Q6_K for half of the attention.wv and feed_forward.w2 tensors, else GGML_TYPE_Q4_K |
+| gpt4-alpaca-lora_mlp-65B.ggmlv3.q4_K_S.bin | q4_K_S | 4 | 36.73 GB | 39.23 GB | New k-quant method. Uses GGML_TYPE_Q4_K for all tensors |
+| gpt4-alpaca-lora_mlp-65B.ggmlv3.q5_0.bin | q5_0 | 5 | 44.89 GB | 47.39 GB | Original llama.cpp quant method, 5-bit. Higher accuracy, higher resource usage and slower inference. |
+| gpt4-alpaca-lora_mlp-65B.ggmlv3.q5_1.bin | q5_1 | 5 | 48.97 GB | 51.47 GB | Original llama.cpp quant method, 5-bit. Even higher accuracy, resource usage and slower inference. |
+| gpt4-alpaca-lora_mlp-65B.ggmlv3.q5_K_M.bin | q5_K_M | 5 | 46.20 GB | 48.70 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 |
+| gpt4-alpaca-lora_mlp-65B.ggmlv3.q5_K_S.bin | q5_K_S | 5 | 44.89 GB | 47.39 GB | New k-quant method. Uses GGML_TYPE_Q5_K for all tensors |
+
+
+**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.
 
-Note: no q8_0 will be provided as HF won't allow uploading of files larger than 50GB :)
+## How to run in `llama.cpp`
 
-I am investigating other methods, eg a split ZIP file, and will try to upload this soon.
+I use the following command line; adjust for your tastes and needs:
+
+```
+./main -t 10 -ngl 32 -m gpt4-alpaca-lora_mlp-65B.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:"
+```
+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`.
+
+Change `-ngl 32` to the number of layers to offload to GPU. Remove it if you don't have GPU acceleration.
+
+If you want to have a chat-style conversation, replace the `-p <PROMPT>` argument with `-i -ins`
+
+## How to run in `text-generation-webui`
+
+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).
 
 <!-- footer start -->
 ## Discord
@@ -70,19 +118,24 @@ Donaters will get priority support on any and all AI/LLM/model questions and req
 * Patreon: https://patreon.com/TheBlokeAI
 * Ko-Fi: https://ko-fi.com/TheBlokeAI
 
-**Patreon special mentions**: Aemon Algiz, Dmitriy Samsonov, Nathan LeClaire, Trenton Dambrowitz, Mano Prime, David Flickinger, vamX, Nikolai Manek, senxiiz, Khalefa Al-Ahmad, Illia Dulskyi, Jonathan Leane, Talal Aujan, V. Lukas, Joseph William Delisle, Pyrater, Oscar Rangel, Lone Striker, Luke Pendergrass, Eugene Pentland, Sebastain Graf, Johann-Peter Hartman.
+**Special thanks to**: Luke from CarbonQuill, Aemon Algiz, Dmitriy Samsonov.
+
+**Patreon special mentions**: Oscar Rangel, Eugene Pentland, Talal Aujan, Cory Kujawski, Luke, Asp the Wyvern, Ai Maven, Pyrater, Alps Aficionado, senxiiz, Willem Michiel, Junyu Yang, trip7s trip, Sebastain Graf, Joseph William Delisle, Lone Striker, Jonathan Leane, Johann-Peter Hartmann, David Flickinger, Spiking Neurons AB, Kevin Schuppel, Mano Prime, Dmitriy Samsonov, Sean Connelly, Nathan LeClaire, Alain Rossmann, Fen Risland, Derek Yates, Luke Pendergrass, Nikolai Manek, Khalefa Al-Ahmad, Artur Olbinski, John Detwiler, Ajan Kanaga, Imad Khwaja, Trenton Dambrowitz, Kalila, vamX, webtim, Illia Dulskyi.
 
 Thank you to all my generous patrons and donaters!
+
 <!-- footer end -->
-# Original model card
 
-This repo provides the training checkpoint of LLaMA on the alpaca_data_gpt4 dataset via LoRA [MLP] on 8xA100(80G).
+# Original model card: Chtan's GPT4 Alpaca LoRA MLP 65B
+
+
+This repo provides the training checkpoint of LLaMA on the alpaca_data_gpt4 dataset via LoRA [MLP] on 8xA100(80G). 
 
 He et al. 2022 gave an insight that FFN can better utilize modification at larger capacities.
 
 The codes is provided by [tloen/alpaca-lora: Instruct-tune LLaMA on consumer hardware (github.com)](https://github.com/tloen/alpaca-lora).
 
-We modify the running scripts to
+We modify the running scripts to 
 ```bash
 torchrun --nproc_per_node=8 finetune.py \
     --base_model '/cache1/chtan/large_models/llama-hf/llama-65b' \
@@ -106,7 +159,7 @@ torchrun --nproc_per_node=8 finetune.py \
 
 **Instruction**: Tell me about alpacas.
 
-**gpt4-alpaca-lora_mlp-65b**:
+**gpt4-alpaca-lora_mlp-65b**: 
 
 Alpacas are small, domesticated mammals that are closely related to llamas. They are native to the Andes Mountains of South America, primarily in Peru, Bolivia, and Chile. These animals have been domesticated for thousands of years and were used by the Incas for their fleece, meat, and as pack animals.