📦 Docker AI Models Repository

Official model cards and documentation for AI models available on Docker Hub.

---

🚀 Models Overview

DeepCoder Preview

📌 Description:
DeepCoder-14B-Preview is a code reasoning LLM fine-tuned to scale up to long context lengths.

📂 Model File: ai/deepcoder-preview.md
🐳 Docker Hub: docker.io/ai/deepcoder-preview

Source:

• https://huggingface.co/agentica-org/DeepCoder-14B-Preview

---

DeepSeek R1 Distill LLaMA

📌 Description:
Distilled LLaMA by DeepSeek, fast and optimized for real-world tasks.

📂 Model File: ai/deepseek-r1-distill-llama.md
🐳 Docker Hub: docker.io/ai/deepseek-r1-distill-llama

Sources:

• https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Llama-8B
• https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Llama-70B

---

Gemma 3

📌 Description:
Google's latest Gemma, small yet strong for chat and generation.

📂 Model File: ai/gemma3.md
🐳 Docker Hub: docker.io/ai/gemma3

Source:

• https://huggingface.co/google/gemma-3-4b-it

---

Gemma 3 QAT

📌 Description:
Google's latest Gemma, in its QAT (quantization aware trained) variant.

📂 Model File: ai/gemma3-qat.md
🐳 Docker Hub: docker.io/ai/gemma3-qat

---

Gemma 3N

📌 Description:
Efficient multimodal AI for text, image, audio, and video on low-resource devices.

📂 Model File: ai/gemma3n.md
🐳 Docker Hub: docker.io/ai/gemma3n

---

GPT-OSS

📌 Description:
OpenAI's open-weight models designed for powerful reasoning, agentic tasks.

📂 Model File: ai/gpt-oss.md
🐳 Docker Hub: docker.io/ai/gpt-oss

---

Granite Embedding Multilingual

📌 Description:
Granite Embedding Multilingual is a 278 million parameter, encoder‑only XLM‑RoBERTa‑style model.

📂 Model File: ai/granite-embedding-multilingual.md
🐳 Docker Hub: docker.io/ai/granite-embedding-multilingual

---

Llama 3.1

📌 Description:
Meta's LLaMA 3.1: Chat-focused, benchmark-strong, multilingual-ready.

📂 Model File: ai/llama3.1.md
🐳 Docker Hub: docker.io/ai/llama3.1

Source:

• https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct

---

Llama 3.2

📌 Description:
Solid LLaMA 3 update, reliable for coding, chat, and Q&A tasks.

📂 Model File: ai/llama3.2.md
🐳 Docker Hub: docker.io/ai/llama3.2

Sources:

• https://huggingface.co/meta-llama/Llama-3.2-3B-Instruct
• https://huggingface.co/meta-llama/Llama-3.2-1B-Instruct

---

Llama 3.3

📌 Description:
Newest LLaMA 3 release with improved reasoning and generation quality.

📂 Model File: ai/llama3.3.md
🐳 Docker Hub: docker.io/ai/llama3.3

---

Magistral Small 3.2

📌 Description:
24B multimodal instruction model by Mistral AI, tuned for accuracy, tool use & fewer repeats.

📂 Model File: ai/magistral-small-3.2.md
🐳 Docker Hub: docker.io/ai/magistral-small-3.2

---

Mistral

📌 Description:
Efficient open model with top-tier performance and fast inference.

📂 Model File: ai/mistral.md
🐳 Docker Hub: docker.io/ai/mistral

Source:

• https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.3

---

Mistral Nemo

📌 Description:
Mistral fine-tuned via NVIDIA NeMo for smoother enterprise use.

📂 Model File: ai/mistral-nemo.md
🐳 Docker Hub: docker.io/ai/mistral-nemo

Source:

• https://huggingface.co/mistralai/Mistral-Nemo-Instruct-2407

---

mxbai-embed-large

📌 Description:
mxbai-embed-large-v1 is a top English embed model by Mixedbread AI, great for RAG and more.

📂 Model File: ai/mxbai-embed-large.md
🐳 Docker Hub: docker.io/ai/mxbai-embed-large

Source:

• https://huggingface.co/mixedbread-ai/mxbai-embed-large-v1

---

Nomic Embed Text v1.5

📌 Description:
Nomic Embed Text v1 is an open‑source, fully auditable text embedding model.

📂 Model File: ai/nomic-embed-text-v1.5.md
🐳 Docker Hub: docker.io/ai/nomic-embed-text-v1.5

---

Phi-4

📌 Description:
Microsoft's compact model, surprisingly capable at reasoning and code.

📂 Model File: ai/phi4.md
🐳 Docker Hub: docker.io/ai/phi4

Source:

• https://huggingface.co/microsoft/phi-4

---

Qwen 2.5

📌 Description:
Versatile Qwen update with better language skills and wider support.

📂 Model File: ai/qwen2.5.md
🐳 Docker Hub: docker.io/ai/qwen2.5

Source:

• https://huggingface.co/Qwen/Qwen2.5-7B-Instruct

---

Qwen 3

📌 Description:
Qwen3 is the latest Qwen LLM, built for top-tier coding, math, reasoning, and language tasks.

📂 Model File: ai/qwen3.md
🐳 Docker Hub: docker.io/ai/qwen3

---

Qwen 3 Coder

📌 Description:
Qwen3-Coder is Qwen's new series of coding agent models.

📂 Model File: ai/qwen3-coder.md
🐳 Docker Hub: docker.io/ai/qwen3-coder

---

QwQ

📌 Description:
Experimental Qwen variant—lean, fast, and a bit mysterious.

📂 Model File: ai/qwq.md
🐳 Docker Hub: docker.io/ai/qwq

Source:

• https://huggingface.co/Qwen/QwQ-32B

---

Seed OSS

📌 Description:
Designed for reasoning, agent and general capabilities, and versatile developer-friendly features.

📂 Model File: ai/seed-oss.md
🐳 Docker Hub: docker.io/ai/seed-oss

---

SmolLM 2

📌 Description:
Tiny LLM built for speed, edge devices, and local development.

📂 Model File: ai/smollm2.md
🐳 Docker Hub: docker.io/ai/smollm2

Sources:

• https://huggingface.co/HuggingFaceTB/SmolLM2-360M-Instruct
• https://huggingface.co/HuggingFaceTB/SmolLM2-135M-Instruct

---

SmolLM 3

📌 Description:
SmolLM3 is a 3.1B model for efficient on-device use, with strong performance in chat.

📂 Model File: ai/smollm3.md
🐳 Docker Hub: docker.io/ai/smollm3

---

Devstral Small 1.1

📌 Description:
Devstral Small 1.1 is an agentic coding LLM (24B) fine-tuned from Mistral-Small-3.1 with a 128K context window.

📂 Model File: ai/devstral-small.md
🐳 Docker Hub: docker.io/ai/devstral-small

🛠️ Tools

Model Cards CLI

A command-line tool for working with model cards. See tools/model-cards-cli/README.md for details.

Key features:

• Update "Available model variants" tables in model card markdown files
• Inspect model repositories to extract metadata
• Upload model overviews to Docker Hub
• Support for custom namespaces and private repositories

📝 Contributing

To add or update model cards:

1. Create/edit the markdown file in the ai/ directory
2. Use the Model Cards CLI to update variant tables
3. Submit a pull request

📄 License

See LICENSE file for details.