unsloth/DeepSeek-V4.1-Flash-1017-GGUF
AI model · https://huggingface.co/unsloth/DeepSeek-V4.1-Flash-1017-GGUF
First seen by Oriel on .
AI summary
DeepSeek-V4.1-Flash-1017-GGUF is a quantized model release whose technical report claims improved reasoning, coding, agentic, and visual benchmark results with shorter, faster task execution.
- The report shows gains over DeepSeek-V4.1-Flash on every listed benchmark, including Terminal-Bench 4.0 (42.3 vs. 31.2) and ProgramBench (29.1 vs. 20.3).
- The repository recommends 128GB+ RAM for its smallest 1-bit quantization, 256GB for 3-bit, and 320GB for full precision.
- Running the model requires a llama.cpp PR or the Unsloth Desktop app; the repository advertises up to 3x faster inference with Unsloth.
- No Jinja-format chat template is provided; encoding.py and the Rust-based deepseek-recipe toolkit support prompt encoding, tool calls, thinking, images, and streamed responses.
- Recommended settings include a 1M-token context window and max_tokens of at least 256K; the model weights are MIT-licensed.
#large language models #gguf #quantization #llama.cpp
Generated by gpt-5.6-sol. Check the original for details.
Metric trends
Last 30 days · Up to 120 recorded samples per source and metric, with reposts kept separate · First and latest retained
Likes
Total likes reported by the source.
- Hugging Face Models: 356, latest observation ; +249 over observed period since
Downloads (30d)
Rolling 30-day downloads. A change in this value is not new downloads during the observed period.
- Hugging Face Models: 867,526, latest observation ; +859,355 over observed period since
Aliases
| Source | External ID | URL |
|---|---|---|
| Hugging Face Models | unsloth/DeepSeek-V4.1-Flash-1017-GGUF | https://huggingface.co/unsloth/DeepSeek-V4.1-Flash-1017-GGUF Read discussion |