Downloads · 30 days
184
16% of all-time downloads
yusufbaykaloglu/qwen3.5-4b-turkish-sft
qwen3.5-4b-turkish-sft is a text generation model from yusufbaykaloglu. Use it when you need the model to write or continue text. It is set up for transformers. The card lists the license as apache-2.0.
<img src="./model-banner.png" width="400"/
Downloads · 30 days
184
16% of all-time downloads
All-time downloads
1.1K
Public
Parameters
4.7B
18.7 GB on disk
Likes
2
Public
Click a slice to open those files.
.safetensors9.3 GB · 100%
How the weights are stored.
BF164.7B · 100%
From the Hugging Face model README
Qwen3.5-4B üzerine Türkçe SFT (Supervised Fine-Tuning) ile eğitilmiş bir dil modelidir. Eğitim verisi olarak helpsteer3-tr veri setinin edit alt kümesi kullanılmıştır.
| Base Model | Qwen/Qwen3.5-4B |
| Language | Turkish (tr), English (en) |
| Architecture | Qwen3_5ForConditionalGeneration |
| Precision | bfloat16 |
| Context Length | 262,144 tokens |
| Hidden Size | 2560 |
| Layers | 32 (hybrid: linear + full attention) |
| Parameters | 4.5B |
| License | Apache 2.0 |
| Method | LoRA (bf16) via Unsloth |
| Dataset | helpsteer3-tr (edit subset) |
| Train Samples | 13,740 |
| Epochs | 2 |
| Learning Rate | 2e-4 (cosine scheduler, 3% warmup) |
| Batch Size | 8 (gradient accumulation: 2, effective: 16) |
| Max Seq Length | 2048 |
| Optimizer | AdamW 8-bit |
| LoRA Rank / Alpha | 16 / 16 |
| LoRA Target Modules | q_proj, k_proj, v_proj, o_proj, gate_proj, up_proj, down_proj |
| Trainable Parameters | 21.2M / 4.5B (0.47%) |
| Final Loss | 1.0971 |
| GPU | A100 |
| Training Time | ~7.8 hours |
Not: Bu model Qwen3.5 tabanlıdır. Thinking modu varsayılan olarak açıktır. Doğrudan yanıt almak için
enable_thinking=Falsekullanın.
from transformers import Qwen3_5ForConditionalGeneration, AutoProcessor
model_id = "yusufbaykaloglu/qwen3.5-4b-turkish-sft"
model = Qwen3_5ForConditionalGeneration.from_pretrained(
model_id,
torch_dtype="auto",
device_map="auto",
)
processor = AutoProcessor.from_pretrained(model_id)
messages = [
{"role": "user", "content": [{"type": "text", "text": "Python'da bir listeyi nasıl sıralarım?"}]}
]
text = processor.apply_chat_template(
messages, tokenize=False, add_generation_prompt=True, enable_thinking=False
)
inputs = processor(text=[text], return_tensors="pt").to(model.device)
outputs = model.generate(**inputs, max_new_tokens=1024, temperature=0.6, top_p=0.95, top_k=20)
response = processor.batch_decode(outputs[:, inputs.input_ids.shape[-1]:], skip_special_tokens=True)[0]
print(response)
Quantize edilmiş versiyonlar için: yusufbaykaloglu/qwen3.5-4b-turkish-sft-gguf
| Q4_K_M | 2.7 GB |
| Q8_0 | 4.5 GB |
| BF16-mmproj | 676 MB |
@misc{yusufbaykaloglu2026qwen3.5turkish,
title={Qwen3.5-4B Turkish SFT},
author={Yusuf Baykaloglu},
year={2026},
url={https://huggingface.co/yusufbaykaloglu/qwen3.5-4b-turkish-sft}
}