OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Models

Cornucopia Llama Fin Chinese

Cornucopia Llama Fin Chinese is an open-source Chinese financial-language-model project that provides LLaMA fine-tunes and a Python framework for model inference, comparison and fine-tuning. Its repository describes pretraining, supervised fine-tuning, RLHF and quantization workflows, and identifies Hugging Face Transformers as its base.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Cornucopia Llama Fin Chinese — OpenViST project profile in AI Models.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it when you need a Chinese financial-domain model or want a codebase for experimenting with LLaMA fine-tuning on finance instruction data.

Who it is for

  • Developers exploring Chinese financial question-answering or text-generation models.
  • Researchers and practitioners preparing instruction data to fine-tune a LLaMA model.
  • Teams comparing the repository’s published finance-focused LoRA fine-tunes.

Documented uses

  • Run inference with a published Chinese finance LoRA model.
  • Compare outputs from multiple models using the supplied comparison script.
  • Prepare custom instruction data and invoke the fine-tuning script.
  • Review a framework covering pretraining, supervised fine-tuning, RLHF and quantization.
Before you choose

Requirements & limits

The public repository identifies the code license as Apache-2.0. Python 3.9 or later is recommended; setup instructions use Git LFS to download a 7B base model.

  • The listed LoRA weights are built on different base models: decapoda-research/llama-7b-hf and Linly-AI/Chinese-LLaMA-7B.
  • The README reports training memory use of about 40 GB at batch size 64 and 65 GB at batch size 96 on its stated training setup.
  • Custom datasets need to follow the repository’s fin_data.json format.
  • The README identifies data-quality and data-variety limitations in earlier datasets and displays differing model answers, so outputs require review.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Prepare the repository and download a base model

Set up the documented Python environment and download a local 7B base model.

Before you start

  • Python 3.9 or later is recommended.
  • The repository and its requirements file.
  • Git LFS.

Steps

  1. Install dependencies with the command below.

    Documented command
    pip install -r requirements.txt

  2. Initialize Git LFS with the command below.

    Documented command
    git lfs install

  3. Run the command below to download the 7B model locally.

    Documented command
    bash ./base_models/load.sh

Expected outcome

The documented dependency-installation and base-model download steps are complete.

Guide limitations

  • These steps cover setup only; they do not run inference or fine-tuning.
  • The README gives no general hardware guarantee for this setup.
Documentation 1 ↗
Source-based practical guide

Prepare instruction data and start fine-tuning

Use a custom dataset in the repository’s documented format and invoke its fine-tuning script.

Before you start

  • A Python environment; Python 3.9 or later is recommended.
  • Custom instruction data prepared in the format of ./instruction_data/fin_data.json.

Steps

  1. Prepare the custom dataset in the format specified by ./instruction_data/fin_data.json.

  2. Run the command below.

    Documented command
    bash ./scripts/finetune.sh

Expected outcome

The documented fine-tuning script is invoked with custom instruction data prepared in the specified format.

Guide limitations

  • The README does not detail additional fine-tuning configuration or report a guaranteed result for custom datasets.
  • Training memory use depends on configuration; the README gives examples from an 80 GB A100 setup.
Documentation 1 ↗
COMPARE OPTIONS

Compare options for Model inference and serving

These projects document a common task and share a project role. This is a source-based comparison, not an execution benchmark or a guarantee of interchangeability.

ProjectDocumented purposePlatformAccess and requirements
Cornucopia Llama Fin ChineseRun inference with a published Chinese finance LoRA model.Linux, macOS, WindowsThe public repository identifies the code license as Apache-2.0. Python 3.9 or later is recommended; setup instructions use Git LFS to download a 7B base model.
Pulse Of MotionEstimate a video's physical frame rate from its visual dynamics.LinuxThe repository is public. Its README describes a released checkpoint that downloads from Hugging Face on first inference, with a manual download option. The strongest internal model is not fully released; the README directs interested users to contact the project about access. No license information is provided in the repository evidence.
Neural Image AssessmentTrain the model using AVA images and rating annotations.Cloud, Linux, macOS, Windows
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
17 October 2026
Original website title
jerry1993-tech/Cornucopia-LLaMA-Fin-Chinese · Official homepage title · retained in its source language
How we review sourcesSuggest a correction