|
--- |
|
license: apache-2.0 |
|
inference: false |
|
--- |
|
|
|
**NOTE: This GGML conversion is primarily for use with llama.cpp.** |
|
- PR #896 was used for q4_0. Everything else is latest as of upload time. |
|
- A warning for q4_2 and q4_3: These are WIP. Do not expect any kind of backwards compatibility until they are finalized. |
|
- 7B can be found here: https://huggingface.co/eachadea/ggml-vicuna-7b-1.1 |
|
- **Choosing the right model:** |
|
- `ggml-vicuna-13b-1.1-q4_0` - Fast, lacks in accuracy. |
|
- `ggml-vicuna-13b-1.1-q4_1` - More accurate, lacks in speed. |
|
|
|
- `ggml-vicuna-13b-1.1-q4_2` - Pretty much a better `q4_0`. Similarly fast, but more accurate. |
|
- `ggml-vicuna-13b-1.1-q4_3` - Pretty much a better `q4_1`. More accurate, still pretty slow. |
|
|
|
- `ggml-vicuna-13b-1.0-uncensored` - Available in `q4_2` and `q4_3`, is an uncensored/unfiltered variant of the model. It is based on the previous release and still uses the `### Human:` syntax. Avoid unless you need it. |
|
|
|
--- |
|
|
|
# Vicuna Model Card |
|
|
|
## Model details |
|
|
|
**Model type:** |
|
Vicuna is an open-source chatbot trained by fine-tuning LLaMA on user-shared conversations collected from ShareGPT. |
|
It is an auto-regressive language model, based on the transformer architecture. |
|
|
|
**Model date:** |
|
Vicuna was trained between March 2023 and April 2023. |
|
|
|
**Organizations developing the model:** |
|
The Vicuna team with members from UC Berkeley, CMU, Stanford, and UC San Diego. |
|
|
|
**Paper or resources for more information:** |
|
https://vicuna.lmsys.org/ |
|
|
|
**License:** |
|
Apache License 2.0 |
|
|
|
**Where to send questions or comments about the model:** |
|
https://github.com/lm-sys/FastChat/issues |
|
|
|
## Intended use |
|
**Primary intended uses:** |
|
The primary use of Vicuna is research on large language models and chatbots. |
|
|
|
**Primary intended users:** |
|
The primary intended users of the model are researchers and hobbyists in natural language processing, machine learning, and artificial intelligence. |
|
|
|
## Training dataset |
|
70K conversations collected from ShareGPT.com. |
|
|
|
## Evaluation dataset |
|
A preliminary evaluation of the model quality is conducted by creating a set of 80 diverse questions and utilizing GPT-4 to judge the model outputs. See https://vicuna.lmsys.org/ for more details. |
|
|
|
## Major updates of weights v1.1 |
|
- Refactor the tokenization and separator. In Vicuna v1.1, the separator has been changed from `"###"` to the EOS token `"</s>"`. This change makes it easier to determine the generation stop criteria and enables better compatibility with other libraries. |
|
- Fix the supervised fine-tuning loss computation for better model quality. |
