Configuration of fine-tuning scripts by YAML file
- Dominant language
- Python
- Stars
- 18
- Forks
- 8
- PR merge metrics
- No merged PRs in 30d
Description
**Is your feature request related to a problem? Please describe.**
Currently, all configurations must be provided via CL or hard-coded.
**Describe the solution you'd like**
I'd like to have a config option by a YAML file:
* Should also allow specification of CL arguments. Priority would be: CL arg -> config.yaml -> defaults in code
* Would solve #9, maybe also #10
* Important sub-task: Support (1) different cache lengths per layer and (2) different cache types per layer. The latter would require some extra refactoring, since right now, cache types must be the same
Contributor guide
Research direction
No files or tests are named. Start by locating the fine-tuning scripts and their command-line argument and default handling, then trace cache length and type configuration by layer. Done means YAML settings work, command-line values override YAML, YAML overrides code defaults, and per-layer cache lengths and types are supported.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python, yaml
- Domain
- machine-learning, tooling
- Issue type
- Feature
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 42/100