Specify number of parameters

#23
Files changed (1) hide show
  1. README.md +1 -1
README.md CHANGED
@@ -9,7 +9,7 @@ license: mit
9
 
10
  DialoGPT is a SOTA large-scale pretrained dialogue response generation model for multiturn conversations.
11
  The [human evaluation results](https://github.com/dreasysnail/Dialogpt_dev#human-evaluation) indicate that the response generated from DialoGPT is comparable to human response quality under a single-turn conversation Turing test.
12
- The model is trained on 147M multi-turn dialogue from Reddit discussion thread.
13
 
14
  * Multi-turn generation examples from an interactive environment:
15
 
 
9
 
10
  DialoGPT is a SOTA large-scale pretrained dialogue response generation model for multiturn conversations.
11
  The [human evaluation results](https://github.com/dreasysnail/Dialogpt_dev#human-evaluation) indicate that the response generated from DialoGPT is comparable to human response quality under a single-turn conversation Turing test.
12
+ The model is trained on 147M multi-turn dialogue from Reddit discussion thread, having 354M parameters (354,823,168).
13
 
14
  * Multi-turn generation examples from an interactive environment:
15