|
Download checkpoint/adapter/README.md from nima1/stackcraft-clef-flash-lora: direct link, hf CLI and curl.
- Browser
- Download file 1.08 kB
-
https://huggingface.co/nima1/stackcraft-clef-flash-lora/resolve/main/checkpoint/adapter/README.md
- Command line
-
hf download hf://nima1/stackcraft-clef-flash-lora/checkpoint/adapter/README.md
-
curl -L -o README.md https://huggingface.co/nima1/stackcraft-clef-flash-lora/resolve/main/checkpoint/adapter/README.md
1.08 kB
| base_model: Cloudflare/clef-flash | |
| library_name: peft | |
| tags: | |
| - lora | |
| - stackcraft | |
| # Stackcraft LoRA adapter component | |
| This Stackcraft adapter adapts [Cloudflare/clef-flash](https://huggingface.co/Cloudflare/clef-flash), | |
| released by Cloudflare and pinned at revision `17f0b0ad64efb65d273590632833508766b2aae6`. | |
| The original model card and license are provided in that upstream repository. | |
| This directory is only one component of the Stackcraft checkpoint. Load the | |
| parent checkpoint's `joint_head.safetensors` and this adapter together through | |
| Stackcraft's `load_checkpoint` function. The native decision head contains learned | |
| parameters; loading only the LoRA adapter does not reproduce the trained model. | |
| Use the exact pinned base and the parent `training_config.json` contract. | |
| This component card does not claim improved performance or publication status. | |
| See the release's full model card and evaluation report for measured results, | |
| limitations, training provenance and reproduction commands. A prepared candidate | |
| has not necessarily been selected by validation. | |