huihui-ai/DeepSeek-R1-V3-Fusion

Overview

huihui-ai/DeepSeek-R1-V3-Fusion is a mixed model that combines the strengths of two powerful DeepSeek-based models: deepseek-ai/DeepSeek-R1-0528 and deepseek-ai/DeepSeek-V3-0324

Although it's a simple mix, the model is usable, and no gibberish has appeared.

Model Details

Progression from No Think to Think

The following table and progress bars illustrate the transition from "No Think" to "Think" across four stages (50, 70, 90, 100).

This idea is still in testing and may not be the final result.

Progression Overview

Stage Name Model R1-0528 V3-0324 Description Progress Bar
1 No Think Fusion-5050 50 50 Relies on intuition, lacks deep thinking, reactive [█████ ] 50%
2 Emerging Think Fusion-7030 70 30 Initial analysis, simple reasoning, partially independent [███████ ] 70%
3 Active Think Fusion-9010 90 10 Systematic thinking, deep reflection, highly independent [█████████ ] 90%
4 Full Think R1 100 0 Comprehensive analysis, foresight, innovative [██████████] 100%

Notes

  • Value: Represents the degree of thinking ability, from 50 (lowest) to 100 (highest).
  • Progress Bar: Uses ASCII characters (█) to visually represent completion, with each █ roughly indicating 10%.
  • Description: Summarizes the core characteristics of each stage.

Trend

The progression from No Think to Full Think reflects a shift from reactive behavior to proactive, systematic, and innovative thinking.

Use with ollama

Note: this model requires Ollama 0.9

You can use huihui_ai/ddeepseek-r1:671b-fusion-5050-Q2_K directly

ollama run huihui_ai/deepseek-r1:671b-fusion-5050-Q2_K

huihui_ai/ddeepseek-r1:671b-fusion-7030-Q2_K

ollama run huihui_ai/deepseek-r1:671b-fusion-7030-Q2_K

Donation

If you like it, please click 'like' and follow us for more updates.
You can follow x.com/support_huihui to get the latest model information from huihui.ai.

Your donation helps us continue our further development and improvement, a cup of coffee can do it.
  • bitcoin(BTC):
  bc1qqnkhuchxw0zqjh2ku3lu4hq45hc6gy84uk70ge
Downloads last month
32
GGUF
Model size
671B params
Architecture
deepseek2
Hardware compatibility
Log In to view the estimation

2-bit

Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for huihui-ai/DeepSeek-R1-V3-Fusion-GGUF

Quantized
(21)
this model