--- base_model: BAAI/AquilaChat2-7B-16K inference: false license: other model_creator: Beijing Academy of Artificial Intelligence model_name: Aquilachat2 7B 16K model_type: aquila prompt_template: > System: A chat between a curious human and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the human's questions. Human: {prompt} Assistant: quantized_by: mzwing --- # AquilaChat2 7B 16K - GGUF - Model creator: [Beijing Academy of Artificial Intelligence](https://huggingface.co/BAAI) - Original model: [AquilaChat2 7B 16K](https://huggingface.co/BAAI/AquilaChat2-7B-16K) ## Description This repo contains GGUF format model files for [Beijing Academy of Artificial Intelligence's Aquilachat2 7B 16K](https://huggingface.co/BAAI/AquilaChat2-7B-16K). These files were quantised using hardware kindly provided by [Google Colab](https://colab.research.google.com/)(Free CPU Machine). [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/mzwing/AI-related/blob/master/notebooks/AquilaChat2_7B_16K_GGUF.ipynb) You can also check it out easily in [my GitHub repo](https://github.com/mzwing/AI-related/blob/master/notebooks/AquilaChat2_7B_16K_GGUF.ipynb). ### About GGUF GGUF is a new format introduced by the llama.cpp team on August 21st 2023. It is a replacement for GGML, which is no longer supported by llama.cpp. Here is an incomplate list of clients and libraries that are known to support GGUF: * [llama.cpp](https://github.com/ggerganov/llama.cpp). The source project for GGUF. Offers a CLI and a server option. * [text-generation-webui](https://github.com/oobabooga/text-generation-webui), the most widely used web UI, with many features and powerful extensions. Supports GPU acceleration. * [KoboldCpp](https://github.com/LostRuins/koboldcpp), a fully featured web UI, with GPU accel across all platforms and GPU architectures. Especially good for story telling. * [LM Studio](https://lmstudio.ai/), an easy-to-use and powerful local GUI for Windows and macOS (Silicon), with GPU acceleration. * [LoLLMS Web UI](https://github.com/ParisNeo/lollms-webui), a great web UI with many interesting and unique features, including a full model library for easy model selection. * [Faraday.dev](https://faraday.dev/), an attractive and easy to use character-based chat GUI for Windows and macOS (both Silicon and Intel), with GPU acceleration. * [ctransformers](https://github.com/marella/ctransformers), a Python library with GPU accel, LangChain support, and OpenAI-compatible AI server. * [llama-cpp-python](https://github.com/abetlen/llama-cpp-python), a Python library with GPU accel, LangChain support, and OpenAI-compatible API server. * [candle](https://github.com/huggingface/candle), a Rust ML framework with a focus on performance, including GPU support, and ease of use. * [Nitro](https://nitro.jan.ai/), a fast, lightweight 3mb inference server to supercharge apps with local AI, and OpenAI-compatible API server. ## Repositories available * [2, 3, 4, 5, 6, 8, 16 and 32-bit GGUF models for CPU+GPU inference](https://huggingface.co/mzwing/AquilaChat2-7B-16K-GGUF) * [Beijing Academy of Artificial Intelligence's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/BAAI/AquilaChat2-7B-16K) ## Prompt template: AquilaChat ``` System: A chat between a curious human and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the human's questions. Human: {prompt} Assistant: ``` ## Compatibility These quantised GGUFv2 files are compatible with llama.cpp from August 27th onwards, as of commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221) They are also compatible with many third party UIs and libraries - please see the list at the top of this README. ## Explanation of quantisation methods
Click to see details 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 Refer to the Provided Files table below to see what files use which methods, and how.
## Provided files | Name | Quant method | Bits | Size | Max RAM required | Use case | | ---- | ---- | ---- | ---- | ---- | ----- | | [AquilaChat2-7B-16K.Q2_K.gguf](https://huggingface.co/mzwing/AquilaChat2-7B-16K-GGUF/blob/main/AquilaChat2-7B-16K.Q2_K.gguf) | Q2_K | 2 | 2.86 GB | untested yet | smallest, significant quality loss - not recommended for most purposes | | [AquilaChat2-7B-16K.Q3_K_S.gguf](https://huggingface.co/mzwing/AquilaChat2-7B-16K-GGUF/blob/main/AquilaChat2-7B-16K.Q3_K_S.gguf) | Q3_K_S | 3 | 3.3 GB | untested yet | very small, high quality loss | | [AquilaChat2-7B-16K.Q3_K_M.gguf](https://huggingface.co/mzwing/AquilaChat2-7B-16K-GGUF/blob/main/AquilaChat2-7B-16K.Q3_K_M.gguf) | Q3_K_M | 3 | 3.65 GB | untested yet | very small, high quality loss | | [AquilaChat2-7B-16K.Q3_K_L.gguf](https://huggingface.co/mzwing/AquilaChat2-7B-16K-GGUF/blob/main/AquilaChat2-7B-16K.Q3_K_L.gguf) | Q3_K_L | 3 | 3.95 GB | untested yet | small, substantial quality loss | | [AquilaChat2-7B-16K.Q4_0.gguf](https://huggingface.co/mzwing/AquilaChat2-7B-16K-GGUF/blob/main/AquilaChat2-7B-16K.Q4_0.gguf) | Q4_0 | 4 | 4.22 GB | untested yet | legacy; small, very high quality loss - prefer using Q3_K_M | | [AquilaChat2-7B-16K.Q4_K_S.gguf](https://huggingface.co/mzwing/AquilaChat2-7B-16K-GGUF/blob/main/AquilaChat2-7B-16K.Q4_K_S.gguf) | Q4_K_S | 4 | 4.25 GB | untested yet | small, greater quality loss | | [AquilaChat2-7B-16K.Q4_K_M.gguf](https://huggingface.co/mzwing/AquilaChat2-7B-16K-GGUF/blob/main/AquilaChat2-7B-16K.Q4_K_M.gguf) | Q4_K_M | 4 | 4.47 GB | untested yet | medium, balanced quality - recommended | | [AquilaChat2-7B-16K.Q5_0.gguf](https://huggingface.co/mzwing/AquilaChat2-7B-16K-GGUF/blob/main/AquilaChat2-7B-16K.Q5_0.gguf) | Q5_0 | 5 | 5.08 GB | untested yet | legacy; medium, balanced quality - prefer using Q4_K_M | | [AquilaChat2-7B-16K.Q5_K_S.gguf](https://huggingface.co/mzwing/AquilaChat2-7B-16K-GGUF/blob/main/AquilaChat2-7B-16K.Q5_K_S.gguf) | Q5_K_S | 5 | 5.08 GB | untested yet | large, low quality loss - recommended | | [AquilaChat2-7B-16K.Q5_K_M.gguf](https://huggingface.co/mzwing/AquilaChat2-7B-16K-GGUF/blob/main/AquilaChat2-7B-16K.Q5_K_M.gguf) | Q5_K_M | 5 | 5.21 GB | untested yet | large, very low quality loss - recommended | | [AquilaChat2-7B-16K.Q6_K.gguf](https://huggingface.co/mzwing/AquilaChat2-7B-16K-GGUF/blob/main/AquilaChat2-7B-16K.Q6_K.gguf) | Q6_K | 6 | 5.99 GB | untested yet | very large, extremely low quality loss | | [AquilaChat2-7B-16K.Q8_0.gguf](https://huggingface.co/mzwing/AquilaChat2-7B-16K-GGUF/blob/main/AquilaChat2-7B-16K.Q8_0.gguf) | Q8_0 | 8 | 7.76 GB | untested yet | very large, extremely low quality loss - not recommended | | [AquilaChat2-7B-16K.F16.gguf](https://huggingface.co/mzwing/AquilaChat2-7B-16K-GGUF/blob/main/AquilaChat2-7B-16K.F16.gguf) | F16 | 16 | 14.6 GB | untested yet | extremely large, extremely low quality loss - not recommended | | [AquilaChat2-7B-16K.F32.gguf](https://huggingface.co/mzwing/AquilaChat2-7B-16K-GGUF/blob/main/AquilaChat2-7B-16K.F32.gguf) | F32 | 32 | 29.2 GB | untested yet | extremely large, extremely low quality loss - not recommended | **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 download GGUF files **Note for manual downloaders:** You almost never want to clone the entire repo! Multiple different quantisation formats are provided, and most users only want to pick and download a single file. The following clients/libraries will automatically download models for you, providing a list of available models to choose from: * LM Studio * LoLLMS Web UI * Faraday.dev ### In `text-generation-webui` Under Download Model, you can enter the model repo: `mzwing/AquilaChat2-7B-16K-GGUF`, and below it, a specific filename to download, such as: `AquilaChat2-7B-16K.Q4_K_M.gguf`. Then click Download. ### On the command line, including multiple files at once I recommend using the `huggingface-hub` Python library: ```shell pip3 install huggingface-hub ``` Then you can download any individual model file to the current directory, at high speed, with a command like this: ```shell huggingface-cli download mzwing/AquilaChat2-7B-16K-GGUF AquilaChat2-7B-16K.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False ```
More advanced huggingface-cli download usage You can also download multiple files at once with a pattern: ```shell huggingface-cli download mzwing/AquilaChat2-7B-16K-GGUF --local-dir . --local-dir-use-symlinks False --include='*Q4_K*gguf' ``` For more documentation on downloading with `huggingface-cli`, please see: [HF -> Hub Python Library -> Download files -> Download from the CLI](https://huggingface.co/docs/huggingface_hub/guides/download#download-from-the-cli). To accelerate downloads on fast connections (1Gbit/s or higher), install `hf_transfer`: ```shell pip3 install hf_transfer ``` And set environment variable `HF_HUB_ENABLE_HF_TRANSFER` to `1`: ```shell HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download mzwing/AquilaChat2-7B-16K-GGUF AquilaChat2-7B-16K.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False ``` Windows Command Line users: You can set the environment variable by running `set HF_HUB_ENABLE_HF_TRANSFER=1` before the download command.
## Example `llama.cpp` command Make sure you are using `llama.cpp` from commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221) or later. ```shell ./main -ngl 32 -m AquilaChat2-7B-16K.Q4_K_M.gguf --color -c 2048 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "System: A chat between a curious human and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the human's questions.\nHuman: {prompt}\nAssistant:" ``` Change `-ngl 32` to the number of layers to offload to GPU. Remove it if you don't have GPU acceleration. Change `-c 2048` to the desired sequence length. For extended sequence models - eg 8K, 16K, 32K - the necessary RoPE scaling parameters are read from the GGUF file and set by llama.cpp automatically. If you want to have a chat-style conversation, replace the `-p ` argument with `-i -ins` For other parameters and how to use them, please refer to [the llama.cpp documentation](https://github.com/ggerganov/llama.cpp/blob/master/examples/main/README.md) ## How to run in `text-generation-webui` Further instructions here: [text-generation-webui/docs/llama.cpp.md](https://github.com/oobabooga/text-generation-webui/blob/main/docs/llama.cpp.md). ## How to run from Python code You can use GGUF models from Python using the [llama-cpp-python](https://github.com/abetlen/llama-cpp-python) or [ctransformers](https://github.com/marella/ctransformers) libraries. ### How to load this model in Python code, using ctransformers #### First install the package Run one of the following commands, according to your system: ```shell # Base ctransformers with no GPU acceleration pip install ctransformers # Or with CUDA GPU acceleration pip install ctransformers[cuda] # Or with AMD ROCm GPU acceleration (Linux only) CT_HIPBLAS=1 pip install ctransformers --no-binary ctransformers # Or with Metal GPU acceleration for macOS systems only CT_METAL=1 pip install ctransformers --no-binary ctransformers ``` #### Simple ctransformers example code ```python from ctransformers import AutoModelForCausalLM # Set gpu_layers to the number of layers to offload to GPU. Set to 0 if no GPU acceleration is available on your system. llm = AutoModelForCausalLM.from_pretrained("mzwing/AquilaChat2-7B-16K-GGUF", model_file="AquilaChat2-7B-16K.Q4_K_M.gguf", model_type="aquila", gpu_layers=50) print(llm("AI is going to")) ``` ## How to use with LangChain Here are guides on using llama-cpp-python and ctransformers with LangChain: * [LangChain + llama-cpp-python](https://python.langchain.com/docs/integrations/llms/llamacpp) * [LangChain + ctransformers](https://python.langchain.com/docs/integrations/providers/ctransformers) ## Thanks, and how to contribute Thanks to [Google Colab](https://colab.research.google.com/)! All the quantised models in this repo are done on the awesome platform. Thanks a lot! Thanks to [llama.cpp](https://github.com/ggerganov/llama.cpp)! It inspired me to explore the inspiring AI field, thanks! Thanks to [TheBloke](https://huggingface.co/TheBloke)! Everything in this repo is a reference to him. You are welcome to create a **PullRequest**! Especially for the **RAM Usage**! # Original model card: Beijing Academy of Artificial Intelligence's Aquilachat2 7B 16K ![Aquila_logo](https://huggingface.co/BAAI/AquilaChat2-7B-16K/resolve/main/log.jpeg?download=true)

English | 简体中文

We opensource our **Aquila2** series, now including **Aquila2**, the base language models, namely **Aquila2-7B** and **Aquila2-34B**, as well as **AquilaChat2**, the chat models, namely **AquilaChat2-7B** and **AquilaChat2-34B**, as well as the long-text chat models, namely **AquilaChat2-7B-16k** and **AquilaChat2-34B-16k** The additional details of the Aquila model will be presented in the official technical report. Please stay tuned for updates on official channels. ## Quick Start AquilaChat2-7B-16K（Chat model） ### 1. Inference ```python import torch from transformers import AutoTokenizer, AutoModelForCausalLM from transformers import BitsAndBytesConfig device = torch.device("cuda:0") model_info = "BAAI/AquilaChat2-7B-16K" tokenizer = AutoTokenizer.from_pretrained(model_info, trust_remote_code=True) quantization_config=BitsAndBytesConfig( load_in_4bit=True, bnb_4bit_use_double_quant=True, bnb_4bit_quant_type="nf4", bnb_4bit_compute_dtype=torch.bfloat16, ) model = AutoModelForCausalLM.from_pretrained(model_info, trust_remote_code=True, torch_dtype=torch.float16, # quantization_config=quantization_config, # Uncomment this line for 4bit quantization ) model.eval() model.to(device) text = "请给出10个要到北京旅游的理由。" from predict import predict out = predict(model, text, tokenizer=tokenizer, max_gen_len=200, top_p=0.95, seed=1234, topk=100, temperature=0.9, sft=True, device=device, model_name="AquilaChat2-7B-16K") print(out) ``` ## License Aquila2 series open-source model is licensed under [BAAI Aquila Model Licence Agreement](https://huggingface.co/BAAI/AquilaChat2-7B-16K/blob/main/BAAI-Aquila-Model-License%20-Agreement.pdf)