janhq
/

Jan-v1-edge-gguf

Text Generation

Model card Files Files and versions

bachvudinh commited on Sep 4

Commit

ef40738

·

verified ·

1 Parent(s): 803b276

Update README.md

Files changed (1) hide show

README.md +2 -2

README.md CHANGED Viewed

@@ -30,7 +30,7 @@ Despite having only 1.7B parameters, **Jan-edge** achieves 83% accuracy—nearly
 ### Chat & Instruction Following
-![image/png](https://cdn-uploads.huggingface.co/production/uploads/655e3b59d5c0d3db5359ca3c/XN5tYBIE5CZ8wfpYM4KCe.png)
 Versus Qwen 3 1.7B Thinking, Jan-edge shows a slight degradation on instruction-following and CreativeWriting, while remaining comparable or better on EQBench and recency QA.
@@ -44,7 +44,7 @@ Jan-v1 is optimized for direct integration with the [Jan App](https://jan.ai/).
 **Using vLLM:**
 ```bash
-vllm serve janhq/Jan-edge \
     --host 0.0.0.0 \
     --port 1234 \
     --enable-auto-tool-choice \

 ### Chat & Instruction Following
+![image/png](https://cdn-uploads.huggingface.co/production/uploads/655e3b59d5c0d3db5359ca3c/4yQHrbyJ_poeXpyyogKuu.png)
 Versus Qwen 3 1.7B Thinking, Jan-edge shows a slight degradation on instruction-following and CreativeWriting, while remaining comparable or better on EQBench and recency QA.
 **Using vLLM:**
 ```bash
+vllm serve janhq/Jan-v1-edge \
     --host 0.0.0.0 \
     --port 1234 \
     --enable-auto-tool-choice \