tatsuya4

6 ms·

1. 1.▲
   by tatsuya4 2y ago

   In real-world use cases, it seems more appropriate to use advanced models to generate suitable rule trees or regular expressions for processing HTML → Markdown, rather than directly using a smaller model to handle each HTML instance. The re
2. 2.▲
   by tatsuya4 2y ago

   It appears that no inference provider currently supports the 72B version.
3. 3.▲
   by tatsuya4 2y ago

   ModelBox have a playground for Qwen Math 7B: https://model.box/try/playground/qwen/qwen2-math-7b-instruct 1.5b: https://model.box/try/playground/qwen/qwen2-math-1.5b-instru...
4. 4.▲
   by tatsuya4 2y ago

   Did you use the LLM aggregator like https://model.box ?
5. 5.▲
   by tatsuya4 2y ago

   Just did a quick test in the https://model.box playground, and it looks like the completion length is noticeably shorter than other models (e.g., gpt-4o). However, the response speed meets expectations..