prithivMLmods/Sqweeks-7B-Instruct

Hugging Face
TEXT GENERATIONPricing:Input $0.4 / Cached $0.08 / Output $0.8Concurrent Unit Cost:1Model Size:7.6BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Feb 5, 2025License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Warm

Sqweeks-7B-Instruct is a 7.6 billion parameter instruction-tuned causal language model developed by prithivMLmods, based on the Qwen 2.5 7B architecture. It is specifically fine-tuned on a synthetic dataset to enhance chain-of-thought (CoT) reasoning, logical problem-solving, and context understanding. This model excels at complex reasoning tasks, structured data processing, and multilingual applications, supporting up to 64K tokens context length.

Loading preview...

Sqweeks-7B-Instruct Overview

Sqweeks-7B-Instruct is a 7.6 billion parameter model built upon the Qwen 2.5 7B architecture, developed by prithivMLmods. It has been fine-tuned using a synthetic dataset focused on open-thoughts and general corpus reasoning entries. This specialized training aims to significantly improve the model's chain-of-thought (CoT) reasoning, logical problem-solving, and overall context comprehension, making it a strong candidate for complex analytical tasks.

Key Capabilities

  • Enhanced Reasoning: Demonstrates improved mathematical reasoning, coding proficiency, and structured data processing.
  • Precise Instruction Following: Optimized for generating accurate and structured outputs, including JSON, and capable of producing long texts (over 8K tokens).
  • High Adaptability: Features better role-playing abilities and robustness against varied system prompts.
  • Extended Context Handling: Supports an impressive context window of up to 64K tokens and can generate outputs up to 4K tokens in length.
  • Multilingual Support: Proficient in over 29 languages, including major global languages like Chinese, English, French, Spanish, Portuguese, and German.

Good For

  • Advanced Reasoning: Ideal for tasks requiring logical deduction, multi-step problem-solving, and deep context understanding.
  • Technical Problem-Solving: Suited for mathematical, scientific, and coding-related queries, including code generation and debugging.
  • Structured Data Tasks: Efficiently processes and generates structured data formats like tables and JSON.
  • Global Applications: Its extensive multilingual support makes it suitable for diverse international use cases.
  • Long-Form Content Generation: Capable of producing detailed documents, reports, and instructional materials.

Finetunes
3 models