ggml-vicuna-7b-1.1 / README.md
eachadea's picture
Update README.md
747824d
metadata
license: apache-2.0
inference: true

NOTE: This GGML conversion is primarily for use with llama.cpp.

  • 7B parameters

  • 4-bit quantized

  • Based on version 1.1

  • Used PR "More accurate Q4_0 and Q4_1 quantizations #896" (should be closer in quality to unquantized)

  • Uncensored variant is available, but it's based on version 1.0 (worse quality wise)

  • For q4_2, "Q4_2 ARM #1046" was used. Will update regularly if new changes are made.

  • Choosing between q4_0, q4_1, and q4_2:

    • 4_0 is the fastest. The quality is the poorest.
    • 4_1 is a lot slower. The quality is noticeably better.
    • 4_2 is almost as fast as 4_0 and about as good as 4_1 on Apple Silicon. On Intel/AMD it's hardly better or faster than 4_1.
  • 13B version of this can be found here: https://huggingface.co/eachadea/ggml-vicuna-13b-1.1



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. (48k for the uncensored variant. 22k worth of garbage removed – see https://huggingface.co/datasets/anon8231489123/ShareGPT_Vicuna_unfiltered)

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.