{"slug": "what-is-the-best-llm-for-shenava-koochik-asr", "title": "What is the \"Best LLM for Shenava-Koochik ASR\"", "summary": "Shenava-Koochik, a 114M-parameter Persian streaming ASR model from Reza2kn's Shenava collection, is not the compute bottleneck in Persian speech pipelines; the LLM decision/intent stage is where failures occur, according to an analysis of the model. For Persian intent and decision-making, general multilingual models such as Qwen2.5 and Gemma outperform most small English-first models in Farsi, while PartAI/Dorna-Llama3-8B-Instruct is fine-tuned specifically for Persian instruction-following and reasoning. On Reachy Mini's onboard Pi, an 8B model is likely too heavy for real-time local inference, so smaller Qwen2.5 1.5B/3B models or routing the LLM call to the HF Inference API/Endpoints is recommended while keeping ASR local.", "body_md": "Worth noting Shenava-Koochik itself is a 114M Persian streaming ASR model (from Reza2kn’s Shenava collection), so ASR shouldn’t be the compute bottleneck here, the LLM stage is where “disition” (decision/intent understanding) is failing.\n\nFor Persian-language intent/decision-making specifically, general multilingual models like Qwen2.5 or Gemma tend to do noticeably better in Farsi than most “small” English-first models, that’s usually the actual fix rather than the ASR choice. If you want something Persian-specialized instead, `PartAI/Dorna-Llama3-8B-Instruct` is fine-tuned specifically for Persian instruction-following and reasoning. If you’re running this on Reachy Mini’s onboard Pi, 8B is probably too heavy for real-time local inference though, you’d want a smaller Qwen2.5 (1.5B/3B) or route the LLM call to HF Inference API/Endpoints instead of running it on-device, keeping ASR local and offloading only the decision step.", "url": "https://wpnews.pro/news/what-is-the-best-llm-for-shenava-koochik-asr", "canonical_source": "https://discuss.huggingface.co/t/what-is-the-best-llm-for-shenava-koochik-asr/180530#post_2", "published_at": "2026-09-16 17:28:31+00:00", "updated_at": "2026-09-16 17:42:57.610162+00:00", "lang": "en", "topics": ["large-language-models", "natural-language-processing", "ai-tools", "ai-infrastructure"], "entities": ["Shenava-Koochik", "Reza2kn", "Qwen2.5", "Gemma", "PartAI/Dorna-Llama3-8B-Instruct", "Reachy Mini", "HF Inference API"], "alternates": {"html": "https://wpnews.pro/news/what-is-the-best-llm-for-shenava-koochik-asr", "markdown": "https://wpnews.pro/news/what-is-the-best-llm-for-shenava-koochik-asr.md", "text": "https://wpnews.pro/news/what-is-the-best-llm-for-shenava-koochik-asr.txt", "jsonld": "https://wpnews.pro/news/what-is-the-best-llm-for-shenava-koochik-asr.jsonld"}}