J

Qwopus3.6-27B-v2-MTP-GGUF

Multimodalpor Jackrong·Página del modelo

Fine-tune multimodal de 27B parámetros de Jackrong sobre Qwen3.6-27B en GGUF, con predicción multi-token y visión para tareas de codificación y agentes.

Share:

Modelo base

Qwen/Qwen3.6-27B

Tarjeta del Modelo

</div>

💡 1. Base Model, Training Library & Cooperation

</div>
</div>
</div>

[!WARNING] Community Release Notice: Qwopus3.6-27B-v2-MTP is an experimental community release intended for research, evaluation, and workflow exploration.


🚀 2. MTP Benchmark: Qwen3.6-27B vs Qwopus3.6-27B-v2-MTP

  • Speed: Qwopus3.6-27B-v2-MTP reaches 10.46 overall tokens/sec, compared with 6.29 tokens/sec for Qwen3.6-27B.
  • Latency: total evaluation time drops from 14,901.69s to 6,487.81s, saving 8,413.88s across the full run.
  • Output shape: MTP produces 67,862 completion tokens versus 93,802 from Qwen3.6-27B, giving a more compact overall response profile.

⚙️ 3. Test Environment & Configuration

  • Compute platform: GB10 dedicated server platform.
  • Evaluation format: same local GGUF server stack for both models.
  • llama-server total context: 49152.
  • Temperature / Top-p: 1.0 / 0.95.
  • Max generated tokens: no explicit cap; generation is bounded by the request budget.
  • Request format: /v1/chat/completions with user content as text payload.
Benchmark Summary: Qwen3.6-27B vs Qwopus3.6-27B-v2-MTP
ModelCompletedAvg SpeedOverall T/sCompletion TokensTotal Time
Qwen3.6-27B306.326.2993,80214,901.69s
Qwopus3.6-27B-v2-MTP3010.6610.4667,8626,487.81s
Domain-Level Performance
DomainQuestionsQwen3.6-27B T/sMTP T/sLatency GainQwen3.6-27B TimeMTP TimeToken Delta
Logic56.3310.772.31x38.5 min16.7 min-26.3%
Coding76.2610.272.25x1.52 h40.6 min-27.3%
DevOps66.2910.392.31x47.4 min20.5 min-28.5%
Math86.2911.002.35x1.01 h25.8 min-25.6%
Edge46.488.282.27x10.3 min4.5 min-43.6%

📊 4. Full 30-Question Comparison


🧭 5. Domain Reading


🎯 6. Recommended Use Cases

  • Agentic coding and code review assistance.
  • DevOps runbooks, configuration generation, and incident diagnosis.
  • Multi-step math and probability derivations.
  • Structured reasoning with explicit intermediate logic.
  • Fast constrained output generation where latency matters.

Autor
J
Jackrong
Usuario
Jackrong
Detalles
Descargas292.6K
Me gusta321
AccesoCódigo Abierto
Tareaimage-text-to-text
Tendencia45
Licenciaapache-2.0
Libreríatransformers
Creado21 may 2026
Actualizado2 jun 2026
Ver en Hugging Face
Idiomas
enzhkorujaes
Entiende todo el contexto.

Regístrate para leer casos de estudio completos, acceder a métricas detalladas y recibir todos los reportes.