texdata/Qwen3.6-35B-A3B-Slovenian-LoRA

Qwen3.6-35B-A3B — Slovenian CPT+SFT LoRA adapter

LoRA adapter for Qwen/Qwen3.6-35B-A3B adding Slovenian fluency, knowledge, and en↔sl translation (continued-pretraining + SFT, both baked into this one adapter). Merging it into the base reproduces the full model.

On held-out evals vs the untuned base: Slovenian-LLM-Eval acc_norm 0.623 → 0.654; translation BLEU en→sl 23.8 → 26.3, sl→en 30.9 → 35.0. Reasoning model.

💡 For plug-and-play chat, a GGUF build (LM Studio / llama.cpp) is at texdata/Qwen3.6-35B-A3B-Slovenian-GGUF.

How to run — load in 4-bit

This qwen3_5_moe arch has a broken fp16/bf16 forward in current transformers (zero hidden state → constant logits → garbage). Load in 4-bit (bitsandbytes nf4) — the quantization it was trained under. (vLLM does not run this arch on Blackwell sm_120 GPUs.)

import torch
from transformers import AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig
from peft import PeftModel

bnb = BitsAndBytesConfig(load_in_4bit=True, bnb_4bit_quant_type="nf4",
                         bnb_4bit_compute_dtype=torch.bfloat16, bnb_4bit_use_double_quant=True)
tok  = AutoTokenizer.from_pretrained("Qwen/Qwen3.6-35B-A3B", trust_remote_code=True)
base = AutoModelForCausalLM.from_pretrained("Qwen/Qwen3.6-35B-A3B", quantization_config=bnb,
                                            trust_remote_code=True, device_map="auto")
model = PeftModel.from_pretrained(base, "texdata/Qwen3.6-35B-A3B-Slovenian-LoRA")

# reasoning model: enable_thinking=False for direct answers/translation (else it spends tokens on <think>)
msg  = [{"role": "user", "content": "Prevedi v slovenščino: Good morning!"}]
text = tok.apply_chat_template(msg, add_generation_prompt=True, tokenize=False, enable_thinking=False)
out  = model.generate(**tok(text, return_tensors="pt").to(model.device), max_new_tokens=128)
print(tok.decode(out[0], skip_special_tokens=True))

License & data provenance

Base Qwen/Qwen3.6-35B-A3B is Apache-2.0. Training data is mixed → this repo is license: other:

Data	License
Slovenian Wikipedia (CPT)	CC BY-SA 4.0 (attribution + ShareAlike)
FineWeb2 sl (CPT)	ODC-BY 1.0
cjvt/GaMS-Nemotron-Chat (SFT)	no explicit license — derived from LMSYS-Chat-1M (custom terms) + NVIDIA Nemotron PT
OPUS-100 en–sl (SFT)	unknown / mixed

⚠️ Resolve the LMSYS-Chat-1M and OPUS-100 terms and pick a final license before public or commercial use.