Type something to search...
Text Gen WebUI

Text Gen WebUI

Text Gen WebUI

36.5k 4.9k
04 May, 2024

What is Oobabooga Text Generation WebUI ?

Oobabooga Text Generation WebUI is a Gradio browser interface for Large Language Models. Supports transformers, GPTQ, AWQ, EXL2, llama.cpp (GGUF), Llama models.

Its goal is to become the AUTOMATIC1111/stable-diffusion-webui of text generation.



Oobabooga Text Generation WebUI Features

  • 3 interface modes: default (two columns), notebook, and chat.

  • Multiple model backends: Transformers, llama.cpp (through llama-cpp-python), ExLlamaV2, AutoGPTQ, AutoAWQ, GPTQ-for-LLaMa, CTransformers, QuIP#.

  • Dropdown menu for quickly switching between different models.

  • Large number of extensions (built-in and user-contributed), including Coqui TTS for realistic voice outputs, Whisper STT for voice inputs, translation, multimodal pipelines, vector databases, Stable Diffusion integration, and a lot more. See the wiki and the extensions directory for details.

  • Chat with custom characters.

  • Precise chat templates for instruction-following models, including Llama-2-chat, Alpaca, Vicuna, Mistral.

  • LoRA: train new LoRAs with your own data, load/unload LoRAs on the fly for generation.

  • Transformers library integration: load models in 4-bit or 8-bit precision through bitsandbytes, use llama.cpp with transformers samplers (llamacpp_HF loader), CPU inference in 32-bit precision using PyTorch.

  • OpenAI-compatible API server with Chat and Completions endpoints — see the examples.

Install Oobabooga Text Generation WebUI

  1. Clone or download the repository.

  2. Run the start_linux.sh, start_windows.bat, start_macos.sh, or start_wsl.bat script depending on your OS.

  3. Select your GPU vendor when asked.

  4. Once the installation ends, browse to http://localhost:7860/?__theme=dark.

  5. Have fun!

To restart the web UI in the future, just run the start_ script again. This script creates an installer_files folder where it sets up the project’s requirements. In case you need to reinstall the requirements, you can simply delete that folder and start the web UI again.

The script accepts command-line flags. Alternatively, you can edit the CMD_FLAGS.txt file with a text editor and add your flags there.

To get updates in the future, run update_linux.sh , update_windows.bat , update_macos.sh , or update_wsl.bat .


The script uses Miniconda to set up a Conda environment in the installer_files folder.

If you ever need to install something manually in the installer_files environment, you can launch an interactive shell using the cmd script: cmd_linux.sh , cmd_windows.bat , cmd_macos.sh , or cmd_wsl.bat .

  • There is no need to run any of those scripts (start_, update_, or cmd_) as admin/root.

  • For additional instructions about AMD and WSL setup, consult the documentation.

  • For automated installation, you can use the GPU_CHOICE, USE_CUDA118, LAUNCH_AFTER_INSTALL, and INSTALL_EXTENSIONS environment variables. For instance: GPU_CHOICE=A USE_CUDA118=FALSE LAUNCH_AFTER_INSTALL=FALSE INSTALL_EXTENSIONS=FALSE ./start_linux.sh.

Manual installation using Conda

Recommended if you have some experience with the command-line.

0. Install Conda


On Linux or WSL, it can be automatically installed with these two commands (source):

Terminal window
curl -sL "https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh" > "Miniconda3.sh"
bash Miniconda3.sh

1. Create a new conda environment

Terminal window
conda create -n textgen python=3.11
conda activate textgen

2. Install Pytorch

Linux/WSLNVIDIApip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121
Linux/WSLCPU onlypip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cpu
LinuxAMDpip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/rocm5.6
MacOS + MPSAnypip3 install torch torchvision torchaudio
WindowsNVIDIApip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121
WindowsCPU onlypip3 install torch torchvision torchaudio

The up-to-date commands can be found here: https://pytorch.org/get-started/locally/.

For NVIDIA, you also need to install the CUDA runtime libraries:

Terminal window
conda install -y -c "nvidia/label/cuda-12.1.1" cuda-runtime

If you need nvcc to compile some library manually, replace the command above with

Terminal window
conda install -y -c "nvidia/label/cuda-12.1.1" cuda

3. Install the web UI

Terminal window
git clone https://github.com/oobabooga/text-generation-webui
cd text-generation-webui
pip install -r <requirements file according to table below>

Requirements file to use:

GPUCPUrequirements file to use
NVIDIAhas AVX2requirements.txt
NVIDIAno AVX2requirements_noavx2.txt
AMDhas AVX2requirements_amd.txt
AMDno AVX2requirements_amd_noavx2.txt
CPU onlyhas AVX2requirements_cpu_only.txt
CPU onlyno AVX2requirements_cpu_only_noavx2.txt
AppleApple Siliconrequirements_apple_silicon.txt

Start the web UI

Terminal window
conda activate textgen
cd text-generation-webui
python server.py

Then browse to


AMD GPU on Windows
  1. Use requirements_cpu_only.txt or requirements_cpu_only_noavx2.txt in the command above.

  2. Manually install llama-cpp-python using the appropriate command for your hardware: Installation from PyPI.

  3. Manually install AutoGPTQ: Installation.

    • Perform the from-source installation - there are no prebuilt ROCm packages for Windows.
  1. For Kepler GPUs and older, you will need to install CUDA 11.8 instead of 12:
Terminal window
pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118
conda install -y -c "nvidia/label/cuda-11.8.0" cuda-runtime
  1. bitsandbytes >= 0.39 may not work. In that case, to use --load-in-8bit, you may have to downgrade like this:

    • Linux: pip install bitsandbytes==0.38.1

    • Windows: pip install https://github.com/jllllll/bitsandbytes-windows-webui/raw/main/bitsandbytes-0.38.1-py3-none-any.whl

Manual install

The requirements*.txt above contain various wheels precompiled through GitHub Actions. If you wish to compile things manually, or if you need to because no suitable wheels are available for your hardware, you can use requirements_nowheels.txt and then install your desired loaders manually.

Alternative: Docker

Terminal window
ln -s docker/{nvidia/Dockerfile,docker-compose.yml,.dockerignore} .
cp docker/.env.example .env
# Edit .env and set:
# TORCH_CUDA_ARCH_LIST based on your GPU model
# APP_RUNTIME_GID your host user's group id (run `id -g` in a terminal)
# BUILD_EXTENIONS optionally add comma separated list of extensions to build
docker compose up --build
  • You need to have Docker Compose v2.17 or higher installed. See this guide for instructions.

  • For additional docker files, check out this repository.

Updating the requirements

From time to time, the requirements*.txt change. To update, use these commands:

Terminal window
conda activate textgen
cd text-generation-webui
pip install -r <requirements file that you have used> --upgrade

Basic settings

-h , --helpshow this help message and exit
--multi-userMulti-user mode. Chat histories are not saved or automatically loaded. WARNING: this is likely not safe for sharing publicly.
--character CHARACTERThe name of the character to load in chat mode by default.
--model MODELName of the model to load by default.
--lora LORA [LORA ...]The list of LoRAs to load. If you want to load more than one LoRA, write the names separated by spaces.
--model-dir MODEL_DIRPath to directory with all the models.
--lora-dir LORA_DIRPath to directory with all the loras.
--model-menuShow a model menu in the terminal when the web UI is first launched.
--settings SETTINGS_FILELoad the default interface settings from this yaml file. See settings-template.yaml for an example. If you create a file called settings.yaml , this file will be loaded by default without the need to use the --settings flag.
--extensions EXTENSIONS [EXTENSIONS ...]The list of extensions to load. If you want to load more than one extension, write the names separated by spaces.
--verbosePrint the prompts to the terminal.
--chat-buttonsShow buttons on the chat tab instead of a hover menu.

Model loader

--loader LOADERChoose the model loader manually, otherwise, it will get autodetected. Valid options: Transformers, llama.cpp, llamacpp_HF, ExLlamav2_HF, ExLlamav2, AutoGPTQ, AutoAWQ, GPTQ-for-LLaMa, ctransformers, QuIP#.


--cpuUse the CPU to generate text. Warning: Training on CPU is extremely slow.
--auto-devicesAutomatically split the model across the available GPU(s) and CPU.
--gpu-memory GPU_MEMORY [GPU_MEMORY ...]Maximum GPU memory in GiB to be allocated per GPU. Example: —gpu-memory 10 for a single GPU, —gpu-memory 10 5 for two GPUs. You can also set values in MiB like —gpu-memory 3500MiB.
--cpu-memory CPU_MEMORYMaximum CPU memory in GiB to allocate for offloaded weights. Same as above.
--diskIf the model is too large for your GPU(s) and CPU combined, send the remaining layers to the disk.
--disk-cache-dir DISK_CACHE_DIRDirectory to save the disk cache to. Defaults to “cache”.
--load-in-8bitLoad the model with 8-bit precision (using bitsandbytes).
--bf16Load the model with bfloat16 precision. Requires NVIDIA Ampere GPU.
--no-cacheSet use_cache to False while generating text. This reduces VRAM usage slightly, but it comes at a performance cost.
--trust-remote-codeSet trust_remote_code=True while loading the model. Necessary for some models.
--no_use_fastSet use_fast=False while loading the tokenizer (it’s True by default). Use this if you have any problems related to use_fast.
--use_flash_attention_2Set use_flash_attention_2=True while loading the model.

bitsandbytes 4-bit

⚠️ Requires minimum compute of 7.0 on Windows at the moment.

--load-in-4bitLoad the model with 4-bit precision (using bitsandbytes).
--use_double_quantuse_double_quant for 4-bit.
--compute_dtype COMPUTE_DTYPEcompute dtype for 4-bit. Valid options: bfloat16, float16, float32.
--quant_type QUANT_TYPEquant_type for 4-bit. Valid options: nf4, fp4.


--tensorcoresUse llama-cpp-python compiled with tensor cores support. This increases performance on RTX cards. NVIDIA only.
--n_ctx N_CTXSize of the prompt context.
--threadsNumber of threads to use.
--threads-batch THREADS_BATCHNumber of threads to use for batches/prompt processing.
--no_mul_mat_qDisable the mulmat kernels.
--n_batchMaximum number of prompt tokens to batch together when calling llama_eval.
--no-mmapPrevent mmap from being used.
--mlockForce the system to keep the model in RAM.
--n-gpu-layers N_GPU_LAYERSNumber of layers to offload to the GPU.
--tensor_split TENSOR_SPLITSplit the model across multiple GPUs. Comma-separated list of proportions. Example: 18, 17.
--numaActivate NUMA task allocation for llama.cpp.
--logits_allNeeds to be set for perplexity evaluation to work. Otherwise, ignore it, as it makes prompt processing slower.
--no_offload_kqvDo not offload the K, Q, V to the GPU. This saves VRAM but reduces the performance.
--cache-capacity CACHE_CAPACITYMaximum cache capacity (llama-cpp-python). Examples: 2000MiB, 2GiB. When provided without units, bytes will be assumed.


--gpu-splitComma-separated list of VRAM (in GB) to use per GPU device for model layers. Example: 20, 7, 7.
--max_seq_len MAX_SEQ_LENMaximum sequence length.
--cfg-cacheExLlamav2_HF: Create an additional cache for CFG negative prompts. Necessary to use CFG with that loader.
--no_flash_attnForce flash-attention to not be used.
--cache_8bitUse 8-bit cache to save VRAM.
--num_experts_per_token NUM_EXPERTS_PER_TOKENNumber of experts to use for generation. Applies to MoE models like Mixtral.


--tritonUse triton.
--no_inject_fused_attentionDisable the use of fused attention, which will use less VRAM at the cost of slower inference.
--no_inject_fused_mlpTriton mode only: disable the use of fused MLP, which will use less VRAM at the cost of slower inference.
--no_use_cuda_fp16This can make models faster on some systems.
--desc_actFor models that don’t have a quantize_config.json, this parameter is used to define whether to set desc_act or not in BaseQuantizeConfig.
--disable_exllamaDisable ExLlama kernel, which can improve inference speed on some systems.
--disable_exllamav2Disable ExLlamav2 kernel.


--wbits WBITSLoad a pre-quantized model with specified precision in bits. 2, 3, 4 and 8 are supported.
--model_type MODEL_TYPEModel type of pre-quantized model. Currently LLaMA, OPT, and GPT-J are supported.
--groupsize GROUPSIZEGroup size.
--pre_layer PRE_LAYER [PRE_LAYER ...]The number of layers to allocate to the GPU. Setting this parameter enables CPU offloading for 4-bit models. For multi-gpu, write the numbers separated by spaces, eg --pre_layer 30 60 .
--checkpoint CHECKPOINTThe path to the quantized checkpoint file. If not specified, it will be automatically detected.
--monkey-patchApply the monkey patch for using LoRAs with quantized models.


--model_type MODEL_TYPEModel type of pre-quantized model. Currently gpt2, gptj, gptneox, falcon, llama, mpt, starcoder (gptbigcode), dollyv2, and replit are supported.


--hqq-backendBackend for the HQQ loader. Valid options: PYTORCH, PYTORCH_COMPILE, ATEN.


--deepspeedEnable the use of DeepSpeed ZeRO-3 for inference via the Transformers integration.
--nvme-offload-dir NVME_OFFLOAD_DIRDeepSpeed: Directory to use for ZeRO-3 NVME offloading.
--local_rank LOCAL_RANKDeepSpeed: Optional argument for distributed setups.

RoPE (for llama.cpp, ExLlamaV2, and transformers)

--alpha_value ALPHA_VALUEPositional embeddings alpha factor for NTK RoPE scaling. Use either this or compress_pos_emb , not both.
--rope_freq_base ROPE_FREQ_BASEIf greater than 0, will be used instead of alpha_value. Those two are related by rope_freq_base = 10000 * alpha_value ^ (64 / 63) .
--compress_pos_emb COMPRESS_POS_EMBPositional embeddings compression factor. Should be set to (context length) / (model's original context length) . Equal to 1/rope_freq_scale .


--listenMake the web UI reachable from your local network.
--listen-port LISTEN_PORTThe listening port that the server will use.
--listen-host LISTEN_HOSTThe hostname that the server will use.
--shareCreate a public URL. This is useful for running the web UI on Google Colab or similar.
--auto-launchOpen the web UI in the default browser upon launch.
--gradio-auth USER:PWDSet Gradio authentication password in the format “username:password”. Multiple credentials can also be supplied with “u1:p1, u2:p2, u3:p3”.
--gradio-auth-path GRADIO_AUTH_PATHSet the Gradio authentication file path. The file should contain one or more user:password pairs in the same format as above.
--ssl-keyfile SSL_KEYFILEThe path to the SSL certificate key file.
--ssl-certfile SSL_CERTFILEThe path to the SSL certificate cert file.


--apiEnable the API extension.
--public-apiCreate a public URL for the API using Cloudfare.
--public-api-id PUBLIC_API_IDTunnel ID for named Cloudflare Tunnel. Use together with public-api option.
--api-port API_PORTThe listening port for the API.
--api-key API_KEYAPI authentication key.
--admin-key ADMIN_KEYAPI authentication key for admin tasks like loading and unloading models. If not set, will be the same as —api-key.
--nowebuiDo not launch the Gradio UI. Useful for launching the API in standalone mode.


--multimodal-pipeline PIPELINEThe multimodal pipeline to use. Examples: llava-7b , llava-13b .

Downloading models

Models should be placed in the folder text-generation-webui/models . They are usually downloaded from Hugging Face.

  • GGUF models are a single file and should be placed directly into models. Example:
Terminal window
└── models
└── llama-2-13b-chat.Q4_K_M.gguf
  • The remaining model types (like 16-bit transformers models and GPTQ models) are made of several files and must be placed in a subfolder. Example:
Terminal window
├── models
├── lmsys_vicuna-33b-v1.3
├── config.json
├── generation_config.json
├── pytorch_model-00001-of-00007.bin
├── pytorch_model-00002-of-00007.bin
├── pytorch_model-00003-of-00007.bin
├── pytorch_model-00004-of-00007.bin
├── pytorch_model-00005-of-00007.bin
├── pytorch_model-00006-of-00007.bin
├── pytorch_model-00007-of-00007.bin
├── pytorch_model.bin.index.json
├── special_tokens_map.json
├── tokenizer_config.json
└── tokenizer.model

In both cases, you can use the “Model” tab of the UI to download the model from Hugging Face automatically. It is also possible to download it via the command-line with

Terminal window
python download-model.py organization/model

Run python download-model.py --help to see all the options.