<!-- markdownlint-disable first-line-h1 --> <!-- markdownlint-disable html --> <!-- markdownlint-disable line-length -->

<img alt="English" src="https://img.shields.io/badge/Language-English-blue.svg"> <img alt="Chinese" src="https://img.shields.io/badge/语言-简体中文-red.svg">

DeepSparkHub

<div align="center" style="line-height: 1;"> <a href="https://www.deepspark.org.cn"><img alt="Homepage" src="https://img.shields.io/badge/DeepSpark-Homepage-blue.svg"/></a> <a href="./LICENSE"><img alt="LICENSE" src="https://img.shields.io/badge/license-Apache%202.0-dfd.svg"></a> <a href="https://gitee.com/deep-spark/deepsparkhub/releases/latest"><img alt="Release" src="https://img.shields.io/github/v/release/deep-spark/deepsparkhub?color=ffa"></a> </div> <br>

DeepSparkHub甄选上百个应用算法和模型，覆盖 AI 和通用计算各领域，支持主流市场智能计算场景，包括智慧城市、数字个人、医疗、教育、通信、能源等多个领域。

模型库

大语言模型（LLM）

| Model | Framework | Dataset/Weight | IXUCA SDK | |----------------------------------------------------------------|--------------------|--------------------------|-----------| | Aquila2-34B | Megatron-DeepSpeed | Bookcorpus | 3.4.0 | | Baichuan2-7B | DeepSpeed | baichuan2-7b-base | 3.4.0 | | Bloom-7B1 | Firefly | school_math_0.25M | 3.4.0 | | ChatGLM-6B | DeepSpeed | ADGEN & chatglm-6b | 3.1.0 | | ChatGLM2-6B SFT | DeepSpeed | ADGEN & chatglm2-6b | 3.4.0 | | ChatGLM3-6B | DeepSpeed | ADGEN & chatglm3-6b | 4.2.0 | | DeepSeekMoE | ColossalAI | deepseek-moe-16b-base | 4.2.0 | | DeepSeek-LLM-7B | verl | deepseek-llm-7b-chat | 4.3.0 | | GLM-4 | Torchrun | glm-4-9b-chat | 4.2.0 | | Gemma-2-2B-IT | verl | gemma-2-2b-it | 4.3.0 | | Llama-7B | ColossalAI | llama-7b-hf | 3.1.0 | | Llama2-7B | Megatron-DeepSpeed | Bookcorpus | 3.1.0 | | Llama2-7B RMF | DeepSpeed | Dahoas/rm-static | 3.1.1 | | Llama2-7B RLHF | Megatron-DeepSpeed | llama2-7b&tiny-llama | 3.4.0 | | Llama2-7B SFT | Megatron-DeepSpeed | GPT Small-117M | 3.1.1 | | Llama2-13B | Megatron-DeepSpeed | Bookcorpus | 3.4.0 | | Llama2-34B | Megatron-DeepSpeed | Bookcorpus | 3.4.0 | | Llama3-8B | Megatron-DeepSpeed | Bookcorpus | 4.2.0 | | Llama3-8B | Megatron-LM | GPT Small-117M | 4.3.0 | | Llama3-8B SFT | ColossalAI | school_math_0.25M | 4.2.0 | | Llama3-8B SFT | OpenRLHF | Meta-Llama-3-8B | 4.3.0 | | Llama3-8B PPO | OpenRLHF | Llama-3-8b-sft-mixture | 4.2.0 | | Llama3-8B DPO | OpenRLHF | Llama-3-8b-sft-mixture | 4.2.0 | | Llama3-8B KTO | OpenRLHF | Llama-3-8b-sft-mixture | 4.2.0 | | Llama3-8B DPO | LLaMA-Factory | Meta-Llama-3-8B-Instruct | 4.3.0 | | Llama3-8B Full SFT | LLaMA-Factory | Meta-Llama-3-8B-Instruct | 4.3.0 | | Llama3-8B KTO | LLaMA-Factory | Meta-Llama-3-8B-Instruct | 4.3.0 | | Llama3-8B Pretrain | LLaMA-Factory | Meta-Llama-3-8B-Instruct | 4.3.0 | | Llama3-8B Reward | LLaMA-Factory | Meta-Llama-3-8B-Instruct | 4.3.0 | | Llama3-8B SFT | LLaMA-Factory | Meta-Llama-3-8B-Instruct | 4.3.0 | | Mamba-2 | Megatron-LM | GPT Small-117M | 4.1.1 | | MiniCPM | DeepSpeed | MiniCPM-2B-sft-bf16 | 4.2.0 | | Mixtral 8x7B | Megatron-LM | GPT Small-117M | 4.2.0 | | Mixtral 8x7B | OpenRLHF | Mixtral-8x7B-v0.1 | 4.3.0 | | Phi-3 | Torchrun | Phi-3-mini-4k-instruct | 4.2.0 | | QWen-7B | Firefly | qwen-7b | 3.4.0 | | QWen1.5-7B | Firefly | school_math | 4.2.0 | | QWen1.5-14B | Firefly | school_math | 4.2.0 | | Qwen2-7B | verl | qwen2-7b | 4.3.0 | | Qwen2.5-7B SFT | LLaMA-Factory | qwen2.5-7b | 4.2.0 | | Qwen2.5-1.5B verl | verl | qwen2.5-1.5b | 4.2.0 | | Qwen2.5-7B verl | verl | qwen2.5-7b | 4.2.0 | | Qwen2.5-3B | ColossalAI | qwen2.5-3b | 4.3.0 | | Qwen3-4B Full SFT | LLaMA-Factory | Qwen3-4B-Instruct-2507 | 4.4.0 | | Qwen3-4B SFT | LLaMA-Factory | Qwen3-4B-Instruct-2507 | 4.4.0 | | Qwen3-4B DPO | LLaMA-Factory | Qwen3-4B-Instruct-2507 | 4.4.0 | | Qwen3-4B KTO | LLaMA-Factory | Qwen3-4B-Instruct-2507 | 4.4.0 | | Qwen3-4B Pretrain | LLaMA-Factory | Qwen3-4B-Instruct-2507 | 4.4.0 | | Qwen3-4B Reward | LLaMA-Factory | Qwen3-4B-Instruct-2507 | 4.4.0 | | Qwen3-8B | verl | qwen3-8b | 4.3.0 | | Yi-6B | DeepSpeed | Yi-6B | 4.2.0 | | Yi-1.5-6B | DeepSpeed | Yi-1.5-6B | 4.2.0 | | Yi-VL-6B | LLaMA-Factory | Yi-VL-6B-hf | 4.2.0 |

计算机视觉（CV）

视觉分类

| Model | Framework | Dataset | IXUCA SDK | |--------------------------------------------------------------------------------------|--------------|----------|-----------| | ACmix | PyTorch | ImageNet | 2.2.0 | | ACNet | PyTorch | ImageNet | 3.1.0 | | AlexNet | PyTorch | ImageNet | 2.2.0 | | AlexNet | TensorFlow | ImageNet | 3.1.0 | | BYOL | PyTorch | ImageNet | 3.1.0 | | CBAM | PyTorch | ImageNet | 3.0.0 | | ConvNext | PyTorch | ImageNet | 2.2.0 | | CspDarknet53 | PyTorch | ImageNet | 3.0.0 | | DenseNet | PaddlePaddle | ImageNet | 3.1.0 | | DenseNet | PyTorch | ImageNet | 2.2.0 | | DPN92 | PyTorch | ImageNet | 2.2.0 | | DPN107 | PyTorch | ImageNet | 2.2.0 | | ECA-MobileNetV2 | PyTorch | ImageNet | 2.2.0 | | ECA-ResNet152 | PyTorch | ImageNet | 2.2.0 | | [EfficientNetB0](models/cv/classification/efficie