Skip to content
Language models

Qwen2.5-Coder-3B for RAG engineering: a three-tier fine-tune

1 min readupdated

The gaurav0506/rag-sft-qwen2.5-coder-3b model is the result of supervised fine-tuning (SFT) of the base Qwen2.5-Coder-3B model for Retrieval-Augmented Generation tasks. The authors explored code evolution from the base version through domain-specific continued pre-training (CPT) to full SFT on 3,000 repository examples. The new model completely eliminates infinite generation loops and unhashable array errors, ensuring 100% AST syntactic validity. This tool will be useful for developers of local AI agents and documentation search systems.

Key features

  • 100% Python AST syntactic validity across 10 autonomous search tasks.
  • Complete elimination of infinite recursive code generation loops.
  • Strict adherence to ChatML chat format boundaries and modern SDK standards.
  • Token accuracy boost up to 74.85% compared to the base version.

What you can do

  • Build local hybrid search pipelines with BM25 and RRF ranking.
  • Write queries for Qdrant and vector databases without namespace hallucinations.
  • Implement Python RAG evaluation metrics such as nDCG and MRR.

Example

Use case: ask the model to write a hybrid search module integrating Qdrant and BM25 for a local RAG pipeline.

Read moreDetails
Benchmarks: token accuracy increased from 50.67% to 74.85%.
Hardware: NVIDIA RTX 4070 Laptop (8188 MiB VRAM), local Ollama and Transformers bfloat16.
Training: 3 epochs, 507 SFT steps on 3,000 instruction pairs.
Discuss Telegram 0 54

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 5.7 GB · ready files6.6 GB
  • GGUF Q8_0recommendedweights 3.1 GB · ready files3.9 GB
  • FP8 / INT8weights 2.9 GB3.8 GB
  • 4-bitweights 1.6 GB2.5 GB

Tip: go 8-bit.GGUF Q8_0 is within about a percent of full precision in quality, takes half the memory and runs faster on RTX 40/50 and H100. Full precision only pays off for fine-tuning.

What to get for this model

Massed ComputeOpen Massed Compute

Your own GPU: GGUF Q8_0 — from RTX 2060 · BF16 — from RTX 5060 Ti 8GB

Massed Compute hourly prices, November 2026. The discounted price applies with the coupon at checkout.

Language models: with a 8192-token context.

Estimate: weights plus working memory. Long videos, high resolutions and long contexts need more.

Run it on Massed Compute

A ready script for the machine we advise above: paste it and the model downloads and starts.

  1. 1. Rent A30

    $0.26 an hour with the coupon. Pick the Ubuntu image when you create the machine.

    Open Massed Compute
  2. 2. Paste the script into the machine's terminal

    Connect over SSH (the Massed Compute panel shows the command) or open a terminal on the machine's desktop.

    HF_TOKEN / CIVITAI_TOKENThe script starts with token lines: paste yours between the quotes. The Hugging Face token is for gated models, the Civitai one for Civitai models.

    # gaurav0506/rag-sft-qwen2.5-coder-3b — A30, GGUF Q8_0
    # NVIDIA driver: CUDA 13 needs 580+. If older or missing, the current nvidia-open is installed (keep yours: KEEP_DRIVER=1)
    DRV=$(nvidia-smi --query-gpu=driver_version --format=csv,noheader 2>/dev/null | head -1 | cut -d. -f1)
    case "$DRV" in ""|*[!0-9]*) DRV=0 ;; esac
    if [ "$DRV" -lt 580 ] && [ -z "${KEEP_DRIVER:-}" ]; then
      . /etc/os-release && \
      curl -fsSLO "https://developer.download.nvidia.com/compute/cuda/repos/ubuntu${VERSION_ID//./}/x86_64/cuda-keyring_1.1-1_all.deb" && \
      sudo dpkg -i cuda-keyring_1.1-1_all.deb && rm -f cuda-keyring_1.1-1_all.deb && \
      sudo apt-get update && \
      sudo DEBIAN_FRONTEND=noninteractive apt-get install -y nvidia-open || \
      { echo "Could not install the NVIDIA driver: see the error above. To keep the current one, run with KEEP_DRIVER=1."; exit 1; }
      echo "NVIDIA driver updated. Reboot the machine (sudo reboot) and run this same script again."
      exit 0
    fi
    curl -fsSL https://ollama.com/install.sh | sh
    ollama run hf.co/gaurav0506/rag-sft-qwen2.5-coder-3b:Q8_0
  3. 3. Use it

    The chat opens right in the terminal. Leave with /bye, start again with the same ollama run command.

Stop the machine in the panel when you are done: it bills by the minute while it runs.

How to speed it up

Picked for this model type. Choose your hardware above to see what fits you exactly.

  • FlashAttention 44.0.0b40from November 18, 2026new release

    fastest exact attention on B200

    Kernels for Blackwell (beta). Exact attention without quantisation — fine for everything, LLMs included.

    How to enable: pip install flash-attn-4; vLLM and SGLang pick it up themselves.

    Earlier generations (2)
    • FlashAttention 3 — The Hopper (H100/H200) version with FP8.
    • FlashAttention 2 — The standard for RTX 30/40/50 and A100: exact attention without the memory cost of long contexts.
  • torch.compile2.15.0from October 28, 2026new release

    10–30% for free

    Compiles the model for your GPU. Slower first run, faster afterwards.

    How to enable: model = torch.compile(model); in ComfyUI the TorchCompileModel node.

  • MTP (multi-token prediction)0.34.0from November 16, 2026new release

    1.5–2× generation speed

    If the model has MTP layers (DeepSeek-V3/R1, Qwen3-Next, GLM, MiMo and others), it predicts several tokens per step itself.

    How to enable: vLLM / SGLang: --speculative-config '{"method": "mtp", "num_speculative_tokens": 1}'. Check the model card.

    Earlier generations (2)
    • EAGLE-3 — Speculative decoding with a small draft head. For models without their own MTP layers.
    • EAGLE-2 / Medusa — Earlier draft-head generations. Still work if that is all you find.
  • FP8 KV cache0.34.0from November 16, 2026new release

    2× more context in the same memory

    Keeps the attention cache in 8 bits, so long contexts fit where they did not.

    How to enable: vLLM: --kv-cache-dtype fp8; llama.cpp: -ctk q8_0 -ctv q8_0.

  • vLLM / SGLang0.34.0from November 16, 2026new release

    many times more tokens/s on a server

    Server engines: batching, FlashAttention, FP8 and speculative decoding out of the box.

    How to enable: vllm serve <model> — an OpenAI-compatible API.

    Earlier generations (1)
    • llama.cpp — GGUF models on any GPU, Mac or even CPU; some layers can stay in RAM.

Model family

  1. Qwen2.5-3Bfamily root
  2. Qwen2.5-Coder-3B
  3. rag-sft-qwen2.5-coder-3b

New posts about versions of Qwen2.5-3B (quants, finetunes, LoRAs) will show up under the bell.

Quantised versions 12All on Hugging Face
Fine-tunes 12All on Hugging Face
LoRAs and adapters 2All on Hugging Face
Merges 3All on Hugging Face

About the model

gaurav0506/rag-sft-qwen2.5-coder-3b
Author
gaurav0506
Category
Language models
Released
Posted

On NeyroLearn

Hugging Face

540

Ask AI about this model

Answered by Gemini 4 · the chat is kept only in this tab

Comments