sarimahsan101/qwen2.5-7b-thinking-esp
sarimahsan101/qwen2.5-7b-thinking-esp is a 7.6 billion parameter Qwen2.5-7B-Instruct based model fine-tuned by sarimahsan101 using LoRA. This model is specifically optimized for generating step-by-step, chain-of-thought reasoning in Spanish and French, with a context length of 512 tokens. It excels at providing structured logical explanations and instruction-following with an engaging tone in multilingual contexts.
Loading preview...
Overview
This model, sarimahsan101/qwen2.5-7b-thinking-esp, is a 7.6 billion parameter Qwen2.5-7B-Instruct variant fine-tuned with LoRA for enhanced reasoning. Its core strength lies in generating step-by-step thinking and logical explanations primarily in Spanish and French, while also supporting English. The fine-tuning process utilized curated multilingual reasoning datasets to improve the coherence and depth of its responses.
Key Capabilities
- Generates chain-of-thought reasoning for complex prompts.
- Produces structured, step-by-step answers.
- Handles multilingual prompts across Spanish, French, and English.
- Maintains an engaging and expressive tone in its outputs.
- Designed for efficient inference with low VRAM usage due to 4-bit quantization.
Good For
- Applications requiring detailed, logical explanations in Spanish or French.
- Educational tools that benefit from step-by-step problem-solving.
- Multilingual chatbots needing to provide structured reasoning.
Limitations
- The model has a limited context window of 512 tokens, which may truncate longer reasoning sequences.
- Performance may degrade in highly technical domains or for languages other than ES/FR/EN.
- Chain-of-thought behavior, while learned, may not always be perfectly consistent.