Buckets:
| # Modelfile for Qwen3.8-27B (Q4_K_M) | |
| # Architecture: Qwen3.5 hybrid (linear/full attention) | |
| # Note: On 24GB VRAM, Ollama can't offload KV cache to RAM, so default ctx is 8K. | |
| # Use run-llamacpp.sh for the full 256K context. | |
| FROM ./Qwen3.8-27B-Q4_K_M.gguf | |
| PARAMETER temperature 0.7 | |
| PARAMETER top_p 0.95 | |
| PARAMETER top_k 40 | |
| PARAMETER repeat_penalty 1.0 | |
| PARAMETER num_predict 8192 | |
| PARAMETER num_ctx 8192 | |
| PARAMETER num_gpu 99 | |
| SYSTEM """You are Qwen 3.8, a helpful, harmless, and honest AI assistant created by Alibaba Cloud. | |
| You are a reasoning model. By default you think step-by-step before answering: | |
| - Your reasoning goes inside thinking blocks | |
| - Your final answer goes after response | |
| - Support tool calling via <tool_call><function=name>...</function></tool_call> format | |
| Be clear, accurate, and concise.""" | |
| LICENSE """ | |
| Apache 2.0 | |
| Copyright (c) 2026 Alibaba Cloud | |
| Licensed under the Apache License, Version 2.0. | |
| """ |
Xet Storage Details
- Size:
- 928 Bytes
- Xet hash:
- 5753c156060a7a311ee12afeda496471eda3032fa87732b1c93264e884d24b6a
·
Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.