[GH-ISSUE #9253] mradermacher / deepseek-moe-16b #6031

Closed
opened 2026-04-12 17:22:15 -05:00 by GiteaMirror · 1 comment
Owner

Originally created by @yisulongte on GitHub (Feb 20, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/9253

Please request to add support for the following 2 models to the library.

https://huggingface.co/mradermacher/deepseek-moe-16b-base-i1-GGUF
https://huggingface.co/mradermacher/deepseek-moe-16b-chat-i1-GGUF

Link Type Size/GB Notes
GGUF i1-IQ1_S 5.3 for the desperate
GGUF i1-IQ1_M 5.6 mostly desperate
GGUF i1-IQ2_XXS 6.0  
GGUF i1-IQ2_XS 6.3  
GGUF i1-IQ2_S 6.4  
GGUF i1-IQ2_M 6.7  
GGUF i1-Q2_K 6.8 IQ3_XXS probably better
GGUF i1-Q2_K_S 6.8 very low quality
GGUF i1-IQ3_XXS 7.3 lower quality
GGUF i1-IQ3_XS 7.5  
GGUF i1-IQ3_S 7.9 beats Q3_K*
GGUF i1-Q3_K_S 7.9 IQ3_XS probably better
GGUF i1-IQ3_M 8.0  
GGUF i1-Q3_K_M 8.6 IQ3_S probably better
GGUF i1-Q3_K_L 8.9 IQ3_M probably better
GGUF i1-IQ4_XS 9.0  
GGUF i1-IQ4_NL 9.4 prefer IQ4_XS
GGUF i1-Q4_0 9.4 fast, low quality
GGUF i1-Q4_K_S 10.0 optimal size/speed/quality
GGUF i1-Q4_1 10.4  
GGUF i1-Q4_K_M 11.0 fast, recommended
GGUF i1-Q5_K_S 11.7  
GGUF i1-Q5_K_M 12.5  
GGUF i1-Q6_K 14.8 practically like static Q6_K

thanks!

Originally created by @yisulongte on GitHub (Feb 20, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/9253 Please request to add support for the following 2 models to the library. https://huggingface.co/mradermacher/deepseek-moe-16b-base-i1-GGUF https://huggingface.co/mradermacher/deepseek-moe-16b-chat-i1-GGUF Link | Type | Size/GB | Notes -- | -- | -- | -- GGUF | i1-IQ1_S | 5.3 | for the desperate GGUF | i1-IQ1_M | 5.6 | mostly desperate GGUF | i1-IQ2_XXS | 6.0 |   GGUF | i1-IQ2_XS | 6.3 |   GGUF | i1-IQ2_S | 6.4 |   GGUF | i1-IQ2_M | 6.7 |   GGUF | i1-Q2_K | 6.8 | IQ3_XXS probably better GGUF | i1-Q2_K_S | 6.8 | very low quality GGUF | i1-IQ3_XXS | 7.3 | lower quality GGUF | i1-IQ3_XS | 7.5 |   GGUF | i1-IQ3_S | 7.9 | beats Q3_K* GGUF | i1-Q3_K_S | 7.9 | IQ3_XS probably better GGUF | i1-IQ3_M | 8.0 |   GGUF | i1-Q3_K_M | 8.6 | IQ3_S probably better GGUF | i1-Q3_K_L | 8.9 | IQ3_M probably better GGUF | i1-IQ4_XS | 9.0 |   GGUF | i1-IQ4_NL | 9.4 | prefer IQ4_XS GGUF | i1-Q4_0 | 9.4 | fast, low quality GGUF | i1-Q4_K_S | 10.0 | optimal size/speed/quality GGUF | i1-Q4_1 | 10.4 |   GGUF | i1-Q4_K_M | 11.0 | fast, recommended GGUF | i1-Q5_K_S | 11.7 |   GGUF | i1-Q5_K_M | 12.5 |   GGUF | i1-Q6_K | 14.8 | practically like static Q6_K thanks!
GiteaMirror added the model label 2026-04-12 17:22:15 -05:00
Author
Owner

@rick-github commented on GitHub (Feb 20, 2025):

On the top right of the model page there is a button "Use this model", click on "Ollama", click on "Copy", paste in to a terminal window.

<!-- gh-comment-id:2671769517 --> @rick-github commented on GitHub (Feb 20, 2025): On the top right of the model page there is a button "Use this model", click on "Ollama", click on "Copy", paste in to a terminal window.
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#6031