Choosing the Best Model for Coding with 8-14B Parameters
The user is looking for the best model for fine-tuning on cybersecurity tasks, with 8-14B parameters. They have $300 Google Cloud credit but can't use it for GPUs.
Why it matters
Choosing the right model for coding and cybersecurity tasks is crucial, especially when working with limited computational resources.
Key Points
- 1Evaluating models like Nvidia Nemotron Cascade for coding tasks
- 2Seeking a model that performs well, but not necessarily at 'sonnet grade'
- 3Planning to fine-tune the model for cybersecurity applications
- 4Exploring options to use the $300 Google Cloud credit, despite not being able to access GPUs
Details
The user is interested in finding the best model for coding tasks, specifically with 8-14 billion parameters. They have evaluated some options like the Nvidia Nemotron Cascade, but are unsure if these models are working well enough for their needs. The user is not looking for a 'sonnet grade' model, but rather one that can be fine-tuned for cybersecurity tasks. They have a $300 Google Cloud credit, but are unable to use it for GPU access, so they are exploring options to fine-tune the model without GPU resources.
No comments yet
Be the first to comment