metadata

inference: false
pipeline_tag: text-generation
license: apache-2.0

Yi-9B-exl2

Original model: Yi-9B
Model creator: 01-ai

Quants

4.25bpw-h8 (main)
4.65bpw-h8
5bpw-h8
6bpw-h8
8bpw-h8

Quantization notes

Made with exllamav2 0.0.15 with the default dataset.
This model can be loaded with apps that have exllamav2 loader, such as Text-Generation-WebUI, KoboldAI, ExUI, etc.

Original model card

Building the Next Generation of Open-Source and Bilingual LLMs

🤗 Hugging Face • 🤖 ModelScope • ✡️ WiseModel

👋 Join us 💬 WeChat (Chinese) !

📕 Table of Contents

What is Yi?
- Introduction
- Models
- News
How to use Yi?
- Quick start
  - Choose your path
  - pip
  - docker
  - llama.cpp
  - conda-lock
  - Web demo
- Fine-tuning
- Quantization
- Deployment
- Learning hub
Why Yi?
- Ecosystem
  - Upstream
  - Downstream
    - Serving
    - Quantization
    - Fine-tuning
    - API
- Benchmarks
  - Base model performance
  - Chat model performance
Who can use Yi?
Misc.

What is Yi?

Introduction

🤖 The Yi series models are the next generation of open-source large language models trained from scratch by 01.AI.
🙌 Targeted as a bilingual language model and trained on 3T multilingual corpus, the Yi series models become one of the strongest LLM worldwide, showing promise in language understanding, commonsense reasoning, reading comprehension, and more. For example,
- Yi-34B-Chat model landed in second place (following GPT-4 Turbo), outperforming other LLMs (such as GPT-4, Mixtral, Claude) on the AlpacaEval Leaderboard (based on data available up to January 2024).
- Yi-34B model ranked first among all existing open-source models (such as Falcon-180B, Llama-70B, Claude) in both English and Chinese on various benchmarks, including Hugging Face Open LLM Leaderboard (pre-trained) and C-Eval (based on data available up to November 2023).
- 🙏 (Credits to Llama) Thanks to the Transformer and Llama open-source communities, as they reduce the efforts required to build from scratch and enable the utilization of the same tools within the AI ecosystem.
If you're interested in Yi's adoption of Llama architecture and license usage policy, see Yi's relation with Llama. ⬇️
💡 TL;DR

The Yi series models adopt the same model architecture as Llama but are NOT derivatives of Llama.
- Both Yi and Llama are all based on the Transformer structure, which has been the standard architecture for large language models since 2018.
- Grounded in the Transformer architecture, Llama has become a new cornerstone for the majority of state-of-the-art open-source models due to its excellent stability, reliable convergence, and robust compatibility. This positions Llama as the recognized foundational framework for models including Yi.
- Thanks to the Transformer and Llama architectures, other models can leverage their power, reducing the effort required to build from scratch and enabling the utilization of the same tools within their ecosystems.
- However, the Yi series models are NOT derivatives of Llama, as they do not use Llama's weights.
  - As Llama's structure is employed by the majority of open-source models, the key factors of determining model performance are training datasets, training pipelines, and training infrastructure.
  - Developing in a unique and proprietary way, Yi has independently created its own high-quality training datasets, efficient training pipelines, and robust training infrastructure entirely from the ground up. This effort has led to excellent performance with Yi series models ranking just behind GPT4 and surpassing Llama on the Alpaca Leaderboard in Dec 2023.

[ Back to top ⬆️ ]

News

🎯 2024-03-06: The Yi-9B is open-sourced and available to the public.

Yi-9B stands out as the top performer among a range of similar-sized open-source models (including Mistral-7B, SOLAR-10.7B, Gemma-7B, DeepSeek-Coder-7B-Base-v1.5 and more), particularly excelling in code, math, common-sense reasoning, and reading comprehension.

🎯 2024-01-23: The Yi-VL models, Yi-VL-34B and Yi-VL-6B, are open-sourced and available to the public.

Yi-VL-34B has ranked first among all existing open-source models in the latest benchmarks, including MMMU and CMMMU (based on data available up to January 2024).

🎯 2023-11-23: Chat models are open-sourced and available to the public.

This release contains two chat models based on previously released base models, two 8-bit models quantized by GPTQ, and two 4-bit models quantized by AWQ.

Yi-34B-Chat
Yi-34B-Chat-4bits
Yi-34B-Chat-8bits
Yi-6B-Chat
Yi-6B-Chat-4bits
Yi-6B-Chat-8bits

You can try some of them interactively at:

🔔 2023-11-23: The Yi Series Models Community License Agreement is updated to v2.1.

🔥 2023-11-08: Invited test of Yi-34B chat model.

Application form:

🎯 2023-11-05: The base models, Yi-6B-200K and Yi-34B-200K, are open-sourced and available to the public.

This release contains two base models with the same parameter sizes as the previous release, except that the context window is extended to 200K.

🎯 2023-11-02: The base models, Yi-6B and Yi-34B, are open-sourced and available to the public.

The first public release contains two bilingual (English/Chinese) base models with the parameter sizes of 6B and 34B. Both of them are trained with 4K sequence length and can be extended to 32K during inference time.

[ Back to top ⬆️ ]

Models

Yi models come in multiple sizes and cater to different use cases. You can also fine-tune Yi models to meet your specific requirements.

If you want to deploy Yi models, make sure you meet the software and hardware requirements.

Chat models

Model	Download
Yi-34B-Chat	• 🤗 Hugging Face • 🤖 ModelScope
Yi-34B-Chat-4bits	• 🤗 Hugging Face • 🤖 ModelScope
Yi-34B-Chat-8bits	• 🤗 Hugging Face • 🤖 ModelScope
Yi-6B-Chat	• 🤗 Hugging Face • 🤖 ModelScope
Yi-6B-Chat-4bits	• 🤗 Hugging Face • 🤖 ModelScope
Yi-6B-Chat-8bits	• 🤗 Hugging Face • 🤖 ModelScope

_{^{- 4-bit series models are quantized by AWQ.
- 8-bit series models are quantized by GPTQ
- All quantized models have a low barrier to use since they can be deployed on consumer-grade GPUs (e.g., 3090, 4090).}}

Base models

Model	Download
Yi-34B	• 🤗 Hugging Face • 🤖 ModelScope
Yi-34B-200K	• 🤗 Hugging Face • 🤖 ModelScope
Yi-9B	• 🤗 Hugging Face
Yi-6B	• 🤗 Hugging Face • 🤖 ModelScope
Yi-6B-200K	• 🤗 Hugging Face • 🤖 ModelScope

_{^{- 200k is roughly equivalent to 400,000 Chinese characters.}}

Model info

For chat and base models

Model	Intro	Default context window	Pretrained tokens	Training Data Date
6B series models	They are suitable for personal and academic use.	4K	3T	Up to June 2023
9B model	It is the best at coding and math in the Yi series models.	4K	Yi-9B is continuously trained based on Yi-6B, using 0.8T tokens.	Up to June 2023
34B series models	They are suitable for personal, academic, and commercial (particularly for small and medium-sized enterprises) purposes. It's a cost-effective solution that's affordable and equipped with emergent ability.	4K	3T	Up to June 2023

For chat models
For chat model limitations, see the explanations below. ⬇️
- Hallucination: This refers to the model generating factually incorrect or nonsensical information. With the model's responses being more varied, there's a higher chance of hallucination that are not based on accurate data or logical reasoning.
- Non-determinism in re-generation: When attempting to regenerate or sample responses, inconsistencies in the outcomes may occur. The increased diversity can lead to varying results even under similar input conditions.
- Cumulative Error: This occurs when errors in the model's responses compound over time. As the model generates more diverse responses, the likelihood of small inaccuracies building up into larger errors increases, especially in complex tasks like extended reasoning, mathematical problem-solving, etc.
- To achieve more coherent and consistent responses, it is advisable to adjust generation configuration parameters such as temperature, top_p, or top_k. These adjustments can help in the balance between creativity and coherence in the model's outputs.

[ Back to top ⬆️ ]

How to use Yi?

Quick start
- Choose your path
- pip
- docker
- conda-lock
- llama.cpp
- Web demo
Fine-tuning
Quantization
Deployment
Learning hub

Quick start

Getting up and running with Yi models is simple with multiple choices available.

Choose your path

Select one of the following paths to begin your journey with Yi!

🎯 Deploy Yi locally

If you prefer to deploy Yi models locally,

🙋‍♀️ and you have sufficient resources (for example, NVIDIA A800 80GB), you can choose one of the following methods:
- pip
- Docker
- conda-lock
🙋‍♀️ and you have limited resources (for example, a MacBook Pro), you can use llama.cpp.

🎯 Not to deploy Yi locally

If you prefer not to deploy Yi models locally, you can explore Yi's capabilities using any of the following options.

🙋‍♀️ Run Yi with APIs

If you want to explore more features of Yi, you can adopt one of these methods:

Yi APIs (Yi official)
- Early access has been granted to some applicants. Stay tuned for the next round of access!
Yi APIs (Replicate)

🙋‍♀️ Run Yi in playground

If you want to chat with Yi with more customizable options (e.g., system prompt, temperature, repetition penalty, etc.), you can try one of the following options:

Yi-34B-Chat-Playground (Yi official)
- Access is available through a whitelist. Welcome to apply (fill out a form in English or Chinese).
Yi-34B-Chat-Playground (Replicate)

🙋‍♀️ Chat with Yi

If you want to chat with Yi, you can use one of these online services, which offer a similar user experience:

Yi-34B-Chat (Yi official on Hugging Face)
- No registration is required.
Yi-34B-Chat (Yi official beta)
- Access is available through a whitelist. Welcome to apply (fill out a form in English or Chinese).

[ Back to top ⬆️ ]

Quick start - pip

This tutorial guides you through every step of running Yi-34B-Chat locally on an A800 (80G) and then performing inference.

Step 0: Prerequisites

Make sure Python 3.10 or a later version is installed.
If you want to run other Yi models, see software and hardware requirements.

Step 1: Prepare your environment

To set up the environment and install the required packages, execute the following command.

git clone https://github.com/01-ai/Yi.git
cd yi
pip install -r requirements.txt

Step 2: Download the Yi model

You can download the weights and tokenizer of Yi models from the following sources:

Step 3: Perform inference

You can perform inference with Yi chat or base models as below.

Perform inference with Yi chat model

Create a file named quick_start.py and copy the following content to it.

from transformers import AutoModelForCausalLM, AutoTokenizer

model_path = '<your-model-path>'

tokenizer = AutoTokenizer.from_pretrained(model_path, use_fast=False)

# Since transformers 4.35.0, the GPT-Q/AWQ model can be loaded using AutoModelForCausalLM.
model = AutoModelForCausalLM.from_pretrained(
    model_path,
    device_map="auto",
    torch_dtype='auto'
).eval()

# Prompt content: "hi"
messages = [
    {"role": "user", "content": "hi"}
]

input_ids = tokenizer.apply_chat_template(conversation=messages, tokenize=True, add_generation_prompt=True, return_tensors='pt')
output_ids = model.generate(input_ids.to('cuda'))
response = tokenizer.decode(output_ids[0][input_ids.shape[1]:], skip_special_tokens=True)

# Model response: "Hello! How can I assist you today?"
print(response)

Run quick_start.py.
```
python quick_start.py
```
Then you can see an output similar to the one below. 🥳
```
Hello! How can I assist you today?
```

Perform inference with Yi base model

Yi-34B

The steps are similar to pip - Perform inference with Yi chat model.

You can use the existing file text_generation.py.
```
python demo/text_generation.py  --model <your-model-path>
```
Then you can see an output similar to the one below. 🥳

Output. ⬇️

Prompt: Let me tell you an interesting story about cat Tom and mouse Jerry,

Generation: Let me tell you an interesting story about cat Tom and mouse Jerry, which happened in my childhood. My father had a big house with two cats living inside it to kill mice. One day when I was playing at home alone, I found one of the tomcats lying on his back near our kitchen door, looking very much like he wanted something from us but couldn’t get up because there were too many people around him! He kept trying for several minutes before finally giving up...

Yi-9B

Input

from transformers import AutoModelForCausalLM, AutoTokenizer

MODEL_DIR = "01-ai/Yi-9B"
model = AutoModelForCausalLM.from_pretrained(MODEL_DIR, torch_dtype="auto")
tokenizer = AutoTokenizer.from_pretrained(MODEL_DIR, use_fast=False)

input_text = "# write the quick sort algorithm"
inputs = tokenizer(input_text, return_tensors="pt").to(model.device)
outputs = model.generate(**inputs, max_length=256)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

Output

# write the quick sort algorithm
def quick_sort(arr):
    if len(arr) <= 1:
        return arr
    pivot = arr[len(arr) // 2]
    left = [x for x in arr if x < pivot]
    middle = [x for x in arr if x == pivot]
    right = [x for x in arr if x > pivot]
    return quick_sort(left) + middle + quick_sort(right)

# test the quick sort algorithm
print(quick_sort([3, 6, 8, 10, 1, 2, 1]))

[ Back to top ⬆️ ]

Quick start - Docker

Run Yi-34B-chat locally with Docker: a step-by-step guide. ⬇️

This tutorial guides you through every step of running Yi-34B-Chat on an A800 GPU or 4*4090 locally and then performing inference.

Step 0: Prerequisites

Make sure you've installed Docker and nvidia-container-toolkit.

Step 1: Start Docker

docker run -it --gpus all \
-v <your-model-path>: /models
ghcr.io/01-ai/yi:latest

Alternatively, you can pull the Yi Docker image from registry.lingyiwanwu.com/ci/01-ai/yi:latest.

Step 2: Perform inference

You can perform inference with Yi chat or base models as below.

Perform inference with Yi chat model

The steps are similar to pip - Perform inference with Yi chat model.

Note that the only difference is to set model_path = '<your-model-mount-path>' instead of model_path = '<your-model-path>'.

Perform inference with Yi base model

The steps are similar to pip - Perform inference with Yi base model.

Note that the only difference is to set --model <your-model-mount-path>' instead of model <your-model-path>.

Quick start - conda-lock

You can use conda-lock to generate fully reproducible lock files for conda environments. ⬇️

You can refer to conda-lock.yml for the exact versions of the dependencies. Additionally, you can utilize micromamba for installing these dependencies.
To install the dependencies, follow these steps:

Install micromamba by following the instructions available here.
Execute micromamba install -y -n yi -f conda-lock.yml to create a conda environment named yi and install the necessary dependencies.

Quick start - llama.cpp

Run Yi-chat-6B-2bits locally with llama.cpp: a step-by-step guide. ⬇️

This tutorial guides you through every step of running a quantized model (Yi-chat-6B-2bits) locally and then performing inference.

Step 0: Prerequisites
Step 1: Download llama.cpp
Step 2: Download Yi model
Step 3: Perform inference

Step 0: Prerequisites

This tutorial assumes you use a MacBook Pro with 16GB of memory and an Apple M2 Pro chip.
Make sure git-lfs is installed on your machine.

Step 1: Download `llama.cpp`

To clone the llama.cpp repository, run the following command.

git clone git@github.com:ggerganov/llama.cpp.git

Step 2: Download Yi model

2.1 To clone XeIaso/yi-chat-6B-GGUF with just pointers, run the following command.

GIT_LFS_SKIP_SMUDGE=1 git clone https://huggingface.co/XeIaso/yi-chat-6B-GGUF

2.2 To download a quantized Yi model (yi-chat-6b.Q2_K.gguf), run the following command.

git-lfs pull --include yi-chat-6b.Q2_K.gguf

Step 3: Perform inference

To perform inference with the Yi model, you can use one of the following methods.

Method 1: Perform inference in terminal
Method 2: Perform inference in web

Method 1: Perform inference in terminal

To compile llama.cpp using 4 threads and then conduct inference, navigate to the llama.cpp directory, and run the following command.

Tips

Replace /Users/yu/yi-chat-6B-GGUF/yi-chat-6b.Q2_K.gguf with the actual path of your model.

By default, the model operates in completion mode.

For additional output customization options (for example, system prompt, temperature, repetition penalty, etc.), run ./main -h to check detailed descriptions and usage.

make -j4 && ./main -m /Users/yu/yi-chat-6B-GGUF/yi-chat-6b.Q2_K.gguf -p "How do you feed your pet fox? Please answer this question in 6 simple steps:\nStep 1:" -n 384 -e

...

How do you feed your pet fox? Please answer this question in 6 simple steps:

Step 1: Select the appropriate food for your pet fox. You should choose high-quality, balanced prey items that are suitable for their unique dietary needs. These could include live or frozen mice, rats, pigeons, or other small mammals, as well as fresh fruits and vegetables.

Step 2: Feed your pet fox once or twice a day, depending on the species and its individual preferences. Always ensure that they have access to fresh water throughout the day.

Step 3: Provide an appropriate environment for your pet fox. Ensure it has a comfortable place to rest, plenty of space to move around, and opportunities to play and exercise.

Step 4: Socialize your pet with other animals if possible. Interactions with other creatures can help them develop social skills and prevent boredom or stress.

Step 5: Regularly check for signs of illness or discomfort in your fox. Be prepared to provide veterinary care as needed, especially for common issues such as parasites, dental health problems, or infections.

Step 6: Educate yourself about the needs of your pet fox and be aware of any potential risks or concerns that could affect their well-being. Regularly consult with a veterinarian to ensure you are providing the best care.

...

Now you have successfully asked a question to the Yi model and got an answer! 🥳

Method 2: Perform inference in web

To initialize a lightweight and swift chatbot, run the following command.

cd llama.cpp
./server --ctx-size 2048 --host 0.0.0.0 --n-gpu-layers 64 --model /Users/yu/yi-chat-6B-GGUF/yi-chat-6b.Q2_K.gguf

Then you can get an output like this:

...

llama_new_context_with_model: n_ctx      = 2048
llama_new_context_with_model: freq_base  = 5000000.0
llama_new_context_with_model: freq_scale = 1
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M2 Pro
ggml_metal_init: picking default device: Apple M2 Pro
ggml_metal_init: ggml.metallib not found, loading from source
ggml_metal_init: GGML_METAL_PATH_RESOURCES = nil
ggml_metal_init: loading '/Users/yu/llama.cpp/ggml-metal.metal'
ggml_metal_init: GPU name:   Apple M2 Pro
ggml_metal_init: GPU family: MTLGPUFamilyApple8 (1008)
ggml_metal_init: hasUnifiedMemory              = true
ggml_metal_init: recommendedMaxWorkingSetSize  = 11453.25 MB
ggml_metal_init: maxTransferRate               = built-in GPU
ggml_backend_metal_buffer_type_alloc_buffer: allocated buffer, size =   128.00 MiB, ( 2629.44 / 10922.67)
llama_new_context_with_model: KV self size  =  128.00 MiB, K (f16):   64.00 MiB, V (f16):   64.00 MiB
ggml_backend_metal_buffer_type_alloc_buffer: allocated buffer, size =     0.02 MiB, ( 2629.45 / 10922.67)
llama_build_graph: non-view tensors processed: 676/676
llama_new_context_with_model: compute buffer total size = 159.19 MiB
ggml_backend_metal_buffer_type_alloc_buffer: allocated buffer, size =   156.02 MiB, ( 2785.45 / 10922.67)
Available slots:
-> Slot 0 - max context: 2048

llama server listening at http://0.0.0.0:8080

To access the chatbot interface, open your web browser and enter http://0.0.0.0:8080 into the address bar.
Enter a question, such as "How do you feed your pet fox? Please answer this question in 6 simple steps" into the prompt window, and you will receive a corresponding answer.

[ Back to top ⬆️ ]

Web demo

You can build a web UI demo for Yi chat models (note that Yi base models are not supported in this senario).

Step 1: Prepare your environment.

Step 2: Download the Yi model.

Step 3. To start a web service locally, run the following command.

python demo/web_demo.py -c <your-model-path>

You can access the web UI by entering the address provided in the console into your browser.

[ Back to top ⬆️ ]

Fine-tuning

bash finetune/scripts/run_sft_Yi_6b.sh

Once finished, you can compare the finetuned model and the base model with the following command:

bash finetune/scripts/run_eval.sh

For advanced usage (like fine-tuning based on your custom data), see the explanations below. ⬇️

Finetune code for Yi 6B and 34B

Preparation

From Image

By default, we use a small dataset from BAAI/COIG to finetune the base model. You can also prepare your customized dataset in the following jsonl format:

{ "prompt": "Human: Who are you? Assistant:", "chosen": "I'm Yi." }

And then mount them in the container to replace the default ones:

docker run -it \
    -v /path/to/save/finetuned/model/:/finetuned-model \
    -v /path/to/train.jsonl:/yi/finetune/data/train.json \
    -v /path/to/eval.jsonl:/yi/finetune/data/eval.json \
    ghcr.io/01-ai/yi:latest \
    bash finetune/scripts/run_sft_Yi_6b.sh

From Local Server

Make sure you have conda. If not, use

mkdir -p ~/miniconda3
wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh -O ~/miniconda3/miniconda.sh
bash ~/miniconda3/miniconda.sh -b -u -p ~/miniconda3
rm -rf ~/miniconda3/miniconda.sh
~/miniconda3/bin/conda init bash
source ~/.bashrc

Then, create a conda env:

conda create -n dev_env python=3.10 -y
conda activate dev_env
pip install torch==2.0.1 deepspeed==0.10 tensorboard transformers datasets sentencepiece accelerate ray==2.7

Hardware Setup

For the Yi-6B model, a node with 4 GPUs, each has GPU mem larger than 60GB is recommended.

For the Yi-34B model, because the usage of zero-offload technique takes a lot CPU memory, please be careful to limit the GPU numbers in 34B finetune training. Please use CUDA_VISIBLE_DEVICES to limit the GPU number (as shown in scripts/run_sft_Yi_34b.sh).

A typical hardware setup for finetuning 34B model is a node with 8GPUS (limit to 4 in running by CUDA_VISIBLE_DEVICES=0,1,2,3), each has GPU mem larger than 80GB, with total CPU mem larger than 900GB.

Quick Start

Download a LLM-base model to MODEL_PATH (6B and 34B). A typical folder of models is like:

|-- $MODEL_PATH
|   |-- config.json
|   |-- pytorch_model-00001-of-00002.bin
|   |-- pytorch_model-00002-of-00002.bin
|   |-- pytorch_model.bin.index.json
|   |-- tokenizer_config.json
|   |-- tokenizer.model
|   |-- ...

Download a dataset from huggingface to local storage DATA_PATH, e.g. Dahoas/rm-static.

|-- $DATA_PATH
|   |-- data
|   |   |-- train-00000-of-00001-2a1df75c6bce91ab.parquet
|   |   |-- test-00000-of-00001-8c7c51afc6d45980.parquet
|   |-- dataset_infos.json
|   |-- README.md

finetune/yi_example_dataset has example datasets, which are modified from BAAI/COIG

|-- $DATA_PATH
    |--data
        |-- train.jsonl
        |-- eval.jsonl

cd into the scripts folder, copy and paste the script, and run. For example:

cd finetune/scripts

bash run_sft_Yi_6b.sh

For the Yi-6B base model, setting training_debug_steps=20 and num_train_epochs=4 can output a chat model, which takes about 20 minutes.

For the Yi-34B base model, it takes a relatively long time for initialization. Please be patient.

Evaluation

cd finetune/scripts

bash run_eval.sh

Then you'll see the answer from both the base model and the finetuned model.

[ Back to top ⬆️ ]

Quantization

GPT-Q

python quantization/gptq/quant_autogptq.py \
  --model /base_model                      \
  --output_dir /quantized_model            \
  --trust_remote_code

Once finished, you can then evaluate the resulting model as follows:

python quantization/gptq/eval_quantized_model.py \
  --model /quantized_model                       \
  --trust_remote_code

For a more detailed explanation, see the explanations below. ⬇️

GPT-Q quantization

GPT-Q is a PTQ(Post-Training Quantization) method. It's memory saving and provides potential speedups while retaining the accuracy of the model.

Yi models can be GPT-Q quantized without a lot of efforts. We provide a step-by-step tutorial below.

To run GPT-Q, we will use AutoGPTQ and exllama. And the huggingface transformers has integrated optimum and auto-gptq to perform GPTQ quantization on language models.

Do Quantization

The quant_autogptq.py script is provided for you to perform GPT-Q quantization:

python quant_autogptq.py --model /base_model \
    --output_dir /quantized_model --bits 4 --group_size 128 --trust_remote_code

Run Quantized Model

You can run a quantized model using the eval_quantized_model.py:

python eval_quantized_model.py --model /quantized_model --trust_remote_code

AWQ

python quantization/awq/quant_autoawq.py \
  --model /base_model                      \
  --output_dir /quantized_model            \
  --trust_remote_code

Once finished, you can then evaluate the resulting model as follows:

python quantization/awq/eval_quantized_model.py \
  --model /quantized_model                       \
  --trust_remote_code

For detailed explanations, see the explanations below. ⬇️

AWQ quantization

AWQ is a PTQ(Post-Training Quantization) method. It's an efficient and accurate low-bit weight quantization (INT3/4) for LLMs.

Yi models can be AWQ quantized without a lot of efforts. We provide a step-by-step tutorial below.

To run AWQ, we will use AutoAWQ.

Do Quantization

The quant_autoawq.py script is provided for you to perform AWQ quantization:

python quant_autoawq.py --model /base_model \
    --output_dir /quantized_model --bits 4 --group_size 128 --trust_remote_code

Run Quantized Model

You can run a quantized model using the eval_quantized_model.py:

python eval_quantized_model.py --model /quantized_model --trust_remote_code

[ Back to top ⬆️ ]

Deployment

If you want to deploy Yi models, make sure you meet the software and hardware requirements.

Software requirements

Before using Yi quantized models, make sure you've installed the correct software listed below.

Model	Software
Yi 4-bit quantized models	AWQ and CUDA
Yi 8-bit quantized models	GPTQ and CUDA

Hardware requirements

Before deploying Yi in your environment, make sure your hardware meets the following requirements.

Chat models

Model	Minimum VRAM	Recommended GPU Example
Yi-6B-Chat	15 GB	1 x RTX 3090 1 x RTX 4090 A10 A30
Yi-6B-Chat-4bits	4 GB	1 x RTX 3060 1 x RTX 4060
Yi-6B-Chat-8bits	8 GB	1 x RTX 3070 1 x RTX 4060
Yi-34B-Chat	72 GB	4 x RTX 4090 A800 (80GB)
Yi-34B-Chat-4bits	20 GB	1 x RTX 3090 1 x RTX 4090 A10 A30 A100 (40GB)
Yi-34B-Chat-8bits	38 GB	2 x RTX 3090 2 x RTX 4090 A800 (40GB)

Below are detailed minimum VRAM requirements under different batch use cases.

Model	batch=1	batch=4	batch=16	batch=32
Yi-6B-Chat	12 GB	13 GB	15 GB	18 GB
Yi-6B-Chat-4bits	4 GB	5 GB	7 GB	10 GB
Yi-6B-Chat-8bits	7 GB	8 GB	10 GB	14 GB
Yi-34B-Chat	65 GB	68 GB	76 GB	> 80 GB
Yi-34B-Chat-4bits	19 GB	20 GB	30 GB	40 GB
Yi-34B-Chat-8bits	35 GB	37 GB	46 GB	58 GB

Base models

Model	Minimum VRAM	Recommended GPU Example
Yi-6B	15 GB	1 x RTX 3090 1 x RTX 4090 A10 A30
Yi-6B-200K	50 GB	A800 (80 GB)
Yi-9B	20 GB	1 x RTX 4090 (24 GB)
Yi-34B	72 GB	4 x RTX 4090 A800 (80 GB)
Yi-34B-200K	200 GB	4 x A800 (80 GB)

[ Back to top ⬆️ ]

Learning hub

If you want to learn Yi, you can find a wealth of helpful educational resources here. ⬇️

Welcome to the Yi learning hub!

Whether you're a seasoned developer or a newcomer, you can find a wealth of helpful educational resources to enhance your understanding and skills with Yi models, including insightful blog posts, comprehensive video tutorials, hands-on guides, and more.

The content you find here has been generously contributed by knowledgeable Yi experts and passionate enthusiasts. We extend our heartfelt gratitude for your invaluable contributions!

At the same time, we also warmly invite you to join our collaborative effort by contributing to Yi. If you have already made contributions to Yi, please don't hesitate to showcase your remarkable work in the table below.

With all these resources at your fingertips, you're ready to start your exciting journey with Yi. Happy learning! 🥳

Tutorials

English tutorials

Type	Deliverable	Date	Author
Video	Run dolphin-2.2-yi-34b on IoT Devices	2023-11-30	Second State
Blog	Running Yi-34B-Chat locally using LlamaEdge	2023-11-30	Second State
Video	Install Yi 34B Locally - Chinese English Bilingual LLM	2023-11-05	Fahd Mirza
Video	Dolphin Yi 34b - Brand New Foundational Model TESTED	2023-11-27	Matthew Berman

Chinese tutorials

Type	Deliverable	Date	Author
Blog	实测零一万物Yi-VL多模态语言模型：能准确“识图吃瓜”	2024-02-02	苏洋
Blog	本地运行零一万物 34B 大模型，使用 Llama.cpp & 21G 显存	2023-11-26	苏洋
Blog	零一万物模型折腾笔记：官方 Yi-34B 模型基础使用	2023-12-10	苏洋
Blog	CPU 混合推理，非常见大模型量化方案：“二三五六” 位量化方案	2023-12-12	苏洋
Blog	单卡 3 小时训练 Yi-6B 大模型 Agent：基于 Llama Factory 实战	2024-01-22	郑耀威
Blog	零一万物开源Yi-VL多模态大模型，魔搭社区推理&微调最佳实践来啦！	2024-01-26	ModelScope
Video	只需 24G 显存，用 vllm 跑起来 Yi-34B 中英双语大模型	2023-12-28	漆妮妮
Video	Yi-VL-34B 多模态大模型 - 用两张 A40 显卡跑起来	2023-01-28	漆妮妮

Why Yi?

Ecosystem
- Upstream
- Downstream
  - Serving
  - Quantization
  - Fine-tuning
  - API
Benchmarks
- Chat model performance
- Base model performance
  - Yi-34B and Yi-34B-200K
  - Yi-9B

Ecosystem

Yi has a comprehensive ecosystem, offering a range of tools, services, and models to enrich your experiences and maximize productivity.

Upstream
Downstream
- Serving
- Quantization
- Fine-tuning
- API

Upstream

The Yi series models follow the same model architecture as Llama. By choosing Yi, you can leverage existing tools, libraries, and resources within the Llama ecosystem, eliminating the need to create new tools and enhancing development efficiency.

For example, the Yi series models are saved in the format of the Llama model. You can directly use LlamaForCausalLM and LlamaTokenizer to load the model. For more information, see Use the chat model.

from transformers import AutoModelForCausalLM, AutoTokenizer

tokenizer = AutoTokenizer.from_pretrained("01-ai/Yi-34b", use_fast=False)

model = AutoModelForCausalLM.from_pretrained("01-ai/Yi-34b", device_map="auto")

[ Back to top ⬆️ ]

Downstream

💡 Tip

Feel free to create a PR and share the fantastic work you've built using the Yi series models.

To help others quickly understand your work, it is recommended to use the format of <model-name>: <model-intro> + <model-highlights>.

Serving

If you want to get up with Yi in a few minutes, you can use the following services built upon Yi.

Yi-34B-Chat: you can chat with Yi using one of the following platforms:
- Yi-34B-Chat | Hugging Face
- Yi-34B-Chat | Yi Platform: Note that currently it's available through a whitelist. Welcome to apply (fill out a form in English or Chinese) and experience it firsthand!
Yi-6B-Chat (Replicate): you can use this model with more options by setting additional parameters and calling APIs.
ScaleLLM: you can use this service to run Yi models locally with added flexibility and customization.

Quantization

If you have limited computational capabilities, you can use Yi's quantized models as follows.

These quantized models have reduced precision but offer increased efficiency, such as faster inference speed and smaller RAM usage.

Fine-tuning

If you're seeking to explore the diverse capabilities within Yi's thriving family, you can delve into Yi's fine-tuned models as below.

TheBloke Models: this site hosts numerous fine-tuned models derived from various LLMs including Yi.

This is not an exhaustive list for Yi, but to name a few sorted on downloads:
SUSTech/SUS-Chat-34B: this model ranked first among all models below 70B and outperformed the twice larger deepseek-llm-67b-chat. You can check the result on the Open LLM Leaderboard.
OrionStarAI/OrionStar-Yi-34B-Chat-Llama: this model excelled beyond other models (such as GPT-4, Qwen-14B-Chat, Baichuan2-13B-Chat) in C-Eval and CMMLU evaluations on the OpenCompass LLM Leaderboard.
NousResearch/Nous-Capybara-34B: this model is trained with 200K context length and 3 epochs on the Capybara dataset.

API

amazing-openai-api: this tool converts Yi model APIs into the OpenAI API format out of the box.
LlamaEdge: this tool builds an OpenAI-compatible API server for Yi-34B-Chat using a portable Wasm (WebAssembly) file, powered by Rust.

[ Back to top ⬆️ ]

Chat model performance

Yi-34B-Chat model demonstrates exceptional performance, ranking first among all existing open-source models in the benchmarks including MMLU, CMMLU, BBH, GSM8k, and more.

Evaluation methods and challenges. ⬇️

Evaluation methods: we evaluated various benchmarks using both zero-shot and few-shot methods, except for TruthfulQA.
Zero-shot vs. few-shot: in chat models, the zero-shot approach is more commonly employed.
Evaluation strategy: our evaluation strategy involves generating responses while following instructions explicitly or implicitly (such as using few-shot examples). We then isolate relevant answers from the generated text.
Challenges faced: some models are not well-suited to produce output in the specific format required by instructions in few datasets, which leads to suboptimal results.

*: C-Eval results are evaluated on the validation datasets

Base model performance

Yi-34B and Yi-34B-200K

The Yi-34B and Yi-34B-200K models stand out as the top performers among open-source models, especially excelling in MMLU, CMMLU, common-sense reasoning, reading comprehension, and more.

Evaluation methods. ⬇️

Disparity in results: while benchmarking open-source models, a disparity has been noted between results from our pipeline and those reported by public sources like OpenCompass.
Investigation findings: a deeper investigation reveals that variations in prompts, post-processing strategies, and sampling techniques across models may lead to significant outcome differences.
Uniform benchmarking process: our methodology aligns with the original benchmarks—consistent prompts and post-processing strategies are used, and greedy decoding is applied during evaluations without any post-processing for the generated content.
Efforts to retrieve unreported scores: for scores that were not reported by the original authors (including scores reported with different settings), we try to get results with our pipeline.
Extensive model evaluation: to evaluate the model’s capability extensively, we adopted the methodology outlined in Llama2. Specifically, we included PIQA, SIQA, HellaSwag, WinoGrande, ARC, OBQA, and CSQA to assess common sense reasoning. SquAD, QuAC, and BoolQ were incorporated to evaluate reading comprehension.
Special configurations: CSQA was exclusively tested using a 7-shot setup, while all other tests were conducted with a 0-shot configuration. Additionally, we introduced GSM8K (8-shot@1), MATH (4-shot@1), HumanEval (0-shot@1), and MBPP (3-shot@1) under the category "Math & Code".
Falcon-180B caveat: Falcon-180B was not tested on QuAC and OBQA due to technical constraints. Its performance score is an average from other tasks, and considering the generally lower scores of these two tasks, Falcon-180B's capabilities are likely not underestimated.

Yi-9B

Yi-9B is almost the best among a range of similar-sized open-source models (including Mistral-7B, SOLAR-10.7B, Gemma-7B, DeepSeek-Coder-7B-Base-v1.5 and more), particularly excelling in code, math, common-sense reasoning, and reading comprehension.

In terms of overall ability (Mean-All), Yi-9B performs the best among similarly sized open-source models, surpassing DeepSeek-Coder, DeepSeek-Math, Mistral-7B, SOLAR-10.7B, and Gemma-7B.

In terms of coding ability (Mean-Code), Yi-9B's performance is second only to DeepSeek-Coder-7B, surpassing Yi-34B, SOLAR-10.7B, Mistral-7B, and Gemma-7B.

In terms of math ability (Mean-Math), Yi-9B's performance is second only to DeepSeek-Math-7B, surpassing SOLAR-10.7B, Mistral-7B, and Gemma-7B.

In terms of common sense and reasoning ability (Mean-Text), Yi-9B's performance is on par with Mistral-7B, SOLAR-10.7B, and Gemma-7B.

[ Back to top ⬆️ ]

Who can use Yi?

Everyone! 🙌 ✅

The Yi series models are free for personal usage, academic purposes, and commercial use. All usage must adhere to the Yi Series Models Community License Agreement 2.1
For free commercial use, you only need to complete this form to get a Yi Model Commercial License.

[ Back to top ⬆️ ]

Misc.

Acknowledgments

A heartfelt thank you to each of you who have made contributions to the Yi community! You have helped Yi not just a project, but a vibrant, growing home for innovation.

[ Back to top ⬆️ ]

Disclaimer

We use data compliance checking algorithms during the training process, to ensure the compliance of the trained model to the best of our ability. Due to complex data and the diversity of language model usage scenarios, we cannot guarantee that the model will generate correct, and reasonable output in all scenarios. Please be aware that there is still a risk of the model producing problematic outputs. We will not be responsible for any risks and issues resulting from misuse, misguidance, illegal usage, and related misinformation, as well as any associated data security concerns.

[ Back to top ⬆️ ]

License

The source code in this repo is licensed under the Apache 2.0 license. The Yi series models are fully open for academic research and free for commercial use, with automatic permission granted upon application. All usage must adhere to the Yi Series Models Community License Agreement 2.1. For free commercial use, you only need to send an email to get official commercial permission.

[ Back to top ⬆️ ]

Yi-9B-exl2

Quants

Quantization notes

Original model card

Building the Next Generation of Open-Source and Bilingual LLMs

What is Yi?

Introduction

News

Models

Chat models

Base models

Model info

How to use Yi?

Quick start

Choose your path

🎯 Deploy Yi locally

🎯 Not to deploy Yi locally

🙋‍♀️ Run Yi with APIs

🙋‍♀️ Run Yi in playground

🙋‍♀️ Chat with Yi

Quick start - pip

Step 0: Prerequisites

Step 1: Prepare your environment

Step 2: Download the Yi model

Step 3: Perform inference

Perform inference with Yi chat model

Perform inference with Yi base model

Quick start - Docker

Step 0: Prerequisites

Step 1: Start Docker

Step 2: Perform inference

Perform inference with Yi chat model

Perform inference with Yi base model

Quick start - conda-lock

Quick start - llama.cpp

Step 0: Prerequisites

Step 1: Download llama.cpp

Step 2: Download Yi model

Step 3: Perform inference

Method 1: Perform inference in terminal

Tips

Method 2: Perform inference in web

Web demo

Fine-tuning

Finetune code for Yi 6B and 34B

Preparation

From Image

From Local Server

Hardware Setup

Quick Start

Evaluation

Quantization

GPT-Q

GPT-Q quantization

Do Quantization

Run Quantized Model

AWQ

AWQ quantization

Do Quantization

Run Quantized Model

Deployment

Software requirements

Hardware requirements

Chat models

Base models

Learning hub

Tutorials

English tutorials

Chinese tutorials

Why Yi?

Ecosystem

Upstream

Downstream

Serving

Quantization

Fine-tuning

API

Benchmarks

Chat model performance

Base model performance

Step 1: Download `llama.cpp`