Add link to Neuron-optimized version

#13
by badaoui HF Staff - opened
Files changed (1) hide show
  1. README.md +13 -1
README.md CHANGED
@@ -134,4 +134,16 @@ SmolLM2 models primarily understand and generate content in English. They can pr
134
  primaryClass={cs.CL},
135
  url={https://arxiv.org/abs/2502.02737},
136
  }
137
- ```
 
 
 
 
 
 
 
 
 
 
 
 
 
134
  primaryClass={cs.CL},
135
  url={https://arxiv.org/abs/2502.02737},
136
  }
137
+ ```
138
+
139
+ ---
140
+ ## ๐Ÿš€ AWS Neuron Optimized Version Available
141
+
142
+ A Neuron-optimized version of this model is available for improved performance on AWS Inferentia/Trainium instances:
143
+
144
+ **[badaoui/HuggingFaceTB-SmolLM2-135M-Instruct-neuron](https://huggingface.co/badaoui/HuggingFaceTB-SmolLM2-135M-Instruct-neuron)**
145
+
146
+ The Neuron-optimized version provides:
147
+ - Pre-compiled artifacts for faster loading
148
+ - Optimized performance on AWS Neuron devices
149
+ - Same model capabilities with improved inference speed