Instructions to use Theogott/spr-qwen3_5-9b-dora-vramsafe-adapter with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use Theogott/spr-qwen3_5-9b-dora-vramsafe-adapter with PEFT:
from peft import PeftModel from transformers import AutoModelForCausalLM base_model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen3.5-9B") model = PeftModel.from_pretrained(base_model, "Theogott/spr-qwen3_5-9b-dora-vramsafe-adapter") - Notebooks
- Google Colab
- Kaggle
SPR Qwen3.5-9B DoRa VRAM-safe Adapter
Fine-tuned DoRa/LoRA adapter for PowerBASIC-focused SPR command and code generation.
Built from Qwen/Qwen3.5-9B and the curated SPR train/validation sets in this
repository.
Usage
Load on top of the base model with PEFT. Use this stop token for inference:
###STOP###.
Files
adapter_model.safetensorsadapter_config.jsonchat_template.jinjatokenizer.jsontokenizer_config.jsonprocessor_config.json
Training summary
- Config:
config/spr_qwen3_5_9b_dora_vramsafe.yaml - Training sequence: early-stop run, final adapter checkpoint written to
models/checkpoints_spr_qwen3_5_9b_dora_vramsafe/final_adapter/
Intended use
- PowerBASIC syntax assistance and SPR task-generation style
- Runtime inference via transformers/peft pipelines with matching tokenizer and chat template from this artifact
- Downloads last month
- 2