| --- |
| library_name: peft |
| pipeline_tag: intent-classification |
| license: apache-2.0 |
| tags: |
| - peft |
| - lora |
| - dora |
| - lora+ |
| - dora |
| - adalora |
| - qlora |
| - text-classification |
| - intent-classification |
| --- |
| |
| # Intent Classifier Experiments |
|
|
| This Hugging Face repo stores all experiment-time adapter artifacts for the |
| intent-classifier project. |
|
|
| ## Purpose |
|
|
| - Keep every adapter checkpoint produced during fine-tuning. |
| - Preserve per-experiment traceability across versions. |
| - Keep release artifacts separate from experimental artifacts. |
|
|
| ## Layout |
|
|
| - v1.0/ |
| - one folder per experiment run |
| - naming format: |
| - {model}_{technique}_{config}_{dataset_size}_{YYYYMMDD-HHMMSS} |
| |
| Each experiment folder can include: |
| |
| - adapter weights |
| - tokenizer/config files |
| - metadata and training outputs |
| |
| ## Related repositories |
| |
| Training code and release models are maintained separately. |
| |
| - Training code: <https://github.com/kon172verma/intent-classifier> |
| - Inference/benchmarking: <https://github.com/kon172verma/intent-classifier-inference> |
| - Release: <https://huggingface.co/kon172verma/intent-classifier> |
| |
| ## Versioning |
| |
| Experiment folders are grouped by release version (for example, v1.0). |
| When a version is finalized, only selected best models are promoted to the |
| release repo. |
| |
