williamliao/Qwen3-8B-EAGLE3-Speculator-GGUF

Qwen3-8B EAGLE3 Speculator GGUF

This repository contains a GGUF conversion of RedHatAI/Qwen3-8B-speculator.eagle3 for use with llama.cpp EAGLE3 speculative decoding.

This is not a standalone chat model. It is an EAGLE3 draft/speculator model and must be used together with the matching target/verifier model.

• Target model: Qwen/Qwen3-8B
• Speculator source: RedHatAI/Qwen3-8B-speculator.eagle3
• Runtime: llama.cpp with --spec-type draft-eagle3

Files

File	Type	Notes
Qwen3-8B-speculator.eagle3-F16.gguf	EAGLE3 speculator GGUF	Converted from the original RedHatAI safetensors checkpoint

Usage with llama.cpp

Example:

llama-server \
  -m Qwen3-8B-Q4_K_M.gguf \
  -md Qwen3-8B-speculator.eagle3-F16.gguf \
  --spec-type draft-eagle3 \
  --spec-draft-n-max 4 \
  --spec-draft-p-min 0.5 \
  -c 32768 \
  -ngl 99 \
  -fa on

Windows CMD example:

llama-server.exe ^
  -m Qwen3-8B-Q4_K_M.gguf ^
  -md Qwen3-8B-speculator.eagle3-F16.gguf ^
  --spec-type draft-eagle3 ^
  --spec-draft-n-max 4 ^
  --spec-draft-p-min 0.5 ^
  -c 32768 ^
  -ngl 99 ^
  -fa on

Important Notes

This GGUF file is only the draft/speculator model. You still need a compatible GGUF of the target model, such as a GGUF conversion or quantization of Qwen/Qwen3-8B.

Do not use this speculator with unrelated models such as Qwen3-14B, Qwen3-32B, Gemma, Llama, or Qwen3.6 models. EAGLE3 speculators are target-specific.

Tested Configuration

Tested with:

• Runtime: llama.cpp with EAGLE3 support
• Target model: Qwen3-8B GGUF
• Draft model: this EAGLE3 GGUF
• Example settings:
  • --spec-type draft-eagle3
  • --spec-draft-n-max 4
  • --spec-draft-p-min 0.5

Local benchmark observations may vary depending on GPU, quantization, context length, batch size, and prompt type.

Benchmark Notes

In local testing, EAGLE3 showed stronger gains on structured or predictable outputs such as:

• code generation
• JSON output
• repeated pattern generation
• short factual responses

It was less effective on some tasks such as translation, creative writing, and long reasoning, where draft acceptance may be lower.

Conversion

Converted with llama.cpp convert_hf_to_gguf.py using the original speculator repository and the matching target model directory.

Example conversion command:

python convert_hf_to_gguf.py \
  RedHatAI/Qwen3-8B-speculator.eagle3 \
  --outtype f16 \
  --target-model-dir Qwen/Qwen3-8B \
  --outfile Qwen3-8B-speculator.eagle3-F16.gguf

PowerShell example:

python .\convert_hf_to_gguf.py `
  "E:\OLLAMA_MODELS\Qwen3-8B-speculator.eagle3" `
  --outtype f16 `
  --target-model-dir "E:\OLLAMA_MODELS\Qwen3-8B" `
  --outfile "E:\OLLAMA_MODELS\Qwen3-8B-speculator.eagle3-F16.gguf"

Credits

Original EAGLE3 speculator model by RedHatAI:

• RedHatAI/Qwen3-8B-speculator.eagle3

Target model:

• Qwen/Qwen3-8B

GGUF support and runtime:

• ggml-org/llama.cpp

License

This repository is a converted GGUF version of the original speculator model. The original model license and usage terms apply. Please refer to the upstream repositories for full license details.