DCL Qwen3 4B Instruct 2507 LoRA

--- license: apache-2.0 base_model: Qwen/Qwen3-4B-Instruct-2507 pipeline_tag: text-generation library_name: peft language: - en tags: - dcl - capability-language - lora - qlora --- # DCL-Qwen3-4B-Instruct-2507 — LoRA adapter (~132 MB) The LoRA adapter (r=16, α=32, attention+MLP targets) behind [DCL-Qwen3-4B-Instruct-2507](https://huggingface.co/RichWoollcott/DCL-Qwen3-4B-Instruct-2507), a fine-tune of [Qwen/Qwen3-4B-Instruct-2507](https://huggingface.co/Qwen/Qwen3-4B-Instruct-2507) that authors and repairs [DCL](https://github.com/russelleast/Capability-Language) (Russell East's Declarative Capability Language). See the main model card for the evaluation table, the required serving prompt, and limitations. ```python from transformers import AutoModelForCausalLM, AutoTokenizer from peft import PeftModel import torch m = AutoModelForCausalLM.from_pretrained("Qwen/Qwen3-4B-Instruct-2507", dtype=torch.bfloat16, device_map="auto") m = PeftModel.from_pretrained(m, "RichWoollcott/DCL-Qwen3-4B-Instruct-2507-LoRA") tok = AutoTokenizer.from_pretrained("Qwen/Qwen3-4B-Instruct-2507") ``` Note: the adapter was trained against the 4-bit quantised base (QLoRA); applied to the full-precision base as above, behaviour matches the released merged model to within quantisation noise — the frozen-exam receipts were graded on the merged Q4_K_M build. Trained on a private corpus of 507 compiler-verified, fully synthetic rows with Unsloth + TRL (PEFT 0.19.1) — kept private to protect the integrity of the frozen evaluation exam. Apache-2.0; base LICENSE retained.

Source context: 13 downloads · 0 likes · Pipeline text-generation · Library peft · Repo RichWoollcott/DCL-Qwen3-4B-Instruct-2507-LoRA

Source context: 30 downloads · 0 likes · Pipeline text-generation · Library peft · Repo RichWoollcott/DCL-Qwen3-4B-Instruct-2507-LoRA