Views
No views yet
Qwen/Qwen3-0.6B-Base plus a QLoRA adapter trained on math level 4-5 data.Qwen/Qwen3-0.6B-Basefinal adapterW_merged = W_base + (B @ A) * scaling1### System:
2{system_prompt}
3
4### Problem:
5{problem}
6
7### Solution:
81You are a precise mathematical problem solver.
2Follow this exact output contract:
31. Solve the problem step by step with concise reasoning.
42. Use valid LaTeX math notation for mathematical expressions.
53. Preserve LaTeX commands such as \frac{...}{...}, \sqrt{...}, x^{...}, subscripts, equations, and inequalities.
64. Put the final answer on its own last line exactly in this form: Final Answer: \boxed{...}
75. Do not use Markdown code fences.
86. Do not switch to a different answer format.1from transformers import AutoTokenizer, AutoModelForCausalLM
2import torch
3
4repo_id = "zannvznn/qwen3-0.6b-math-l45-qlora-merged-fp16"
5tokenizer = AutoTokenizer.from_pretrained(repo_id, trust_remote_code=True)
6model = AutoModelForCausalLM.from_pretrained(repo_id, torch_dtype=torch.float16, device_map="auto", trust_remote_code=True)