Spaces:

jbilcke-hf
/

template-node-wizardcoder-express

Paused

Julian Bilcke commited on Jun 22, 2023

Commit

36f130e

1 Parent(s): fa9544d

added a clarification about which machine you need

Files changed (2) hide show

README.md CHANGED Viewed

@@ -69,4 +69,9 @@ You will see the following error message because Docker won't be able to use the
 However if you run your project on a Hugging Face space, you should be just fine :)
-See this demo: https://huggingface.co/spaces/jbilcke-hf/template-node-wizardcoder-express

 However if you run your project on a Hugging Face space, you should be just fine :)
+### Deployment to Hugging Face
+The standard free CPU instance (16 Gb) will not be enough for WizardCoder, you should use the upgraded CPU instance (32 Gb)
+I haven't upgraded mine yet, so it will probably crash when you try it:
+https://huggingface.co/spaces/jbilcke-hf/template-node-wizardcoder-express

src/index.mts CHANGED Viewed

@@ -12,9 +12,15 @@ const app = express()
 const port = 7860
 app.get('/', async (req, res) => {
-  const prompt = '<html><head><title>My Favorite Cookie Recipe</title></head><body><div><p>'
-  res.write(prompt)
   const raw = await llm(prompt)
   const output = raw.split('</html>')
   res.write(output + '</html>')
   res.end()

 const port = 7860
 app.get('/', async (req, res) => {
+  const prefix = `<html><head>`
+  const prompt = `# Context
+A website about a delicious chocolate chip cookie recipe!
+# Output
+${prefix}`
+  res.write(prefix)
+  console.log('calling llm')
   const raw = await llm(prompt)
+  console.log('raw:', raw)
   const output = raw.split('</html>')
   res.write(output + '</html>')
   res.end()