silent fail in llama/main

[tobiasgrossmann]

no Error and no call of callback method.

await dalai.request({
  // model: `llama.13B`,
  model: `${app.locals.configuration.model_type}.${app.locals.configuration.model}`,
  prompt: "Hallo World:",
  skip_end: true,
}, (token) => {
  console.debug("token", token);
  process.stdout.write(token);
  res.write(token);
})

output in shell: repo/robot/llama/main --seed -1 --threads 8 --n_predict 128 --model models/13B/ggml-model-q4_0.bin -p "Hallo World:" in repo/robot/llama

the path models/13B/ggml-model-q4_0.bin seems to be correct and files are there.

"dependencies": { "dalai": "^0.3.1", "express": "^4.18.2", "express-compression": "^1.0.2", "helmet": "^6.1.5", "llama": "^1.3.1", "morgan": "^1.10.0", "node-cron": "^3.0.2", "rss-parser": "^3.13.0", "serve": "^14.2.0" }

[tobiasgrossmann]

worked after reboot. Guess some memory issue.