AccelerateScience/Qwen3-14B-bo-press-conference-sft-merged

TEXT GENERATIONConcurrency Cost:1Model Size:14BQuant:FP8Ctx Length:32kTool Calling:SupportedPublished:Jun 25, 2026License:gpl-3.0Architecture:Transformer Open Weights Cold

AccelerateScience/Qwen3-14B-bo-press-conference-sft-merged is a 14 billion parameter language model, merged from an adapter-only fine-tune of the Qwen3 architecture. This model is specifically fine-tuned for a 'press conference' style of interaction, indicated by its 'bo-press-conference-sft' designation. It is optimized for generating responses within this specific conversational context, offering a specialized solution for related applications.

Loading preview...

Model Overview

AccelerateScience/Qwen3-14B-bo-press-conference-sft-merged is a 14 billion parameter language model built upon the Qwen3 architecture. This particular version is a merged iteration of an adapter-only fine-tuned model, specifically designated for 'bo-press-conference-sft' tasks. The model's validation score is reported at 0.030 [0.028, 0.033], indicating its performance within its specialized domain.

Key Characteristics

  • Architecture: Based on the Qwen3 model family.
  • Parameter Count: 14 billion parameters.
  • Context Length: Supports a context length of 32768 tokens.
  • Fine-tuning: Specialized fine-tuning for 'press conference' style interactions.
  • Generation Configuration: Utilizes a do_sample: True setting with temperature: 0.7, top_p: 1.0, top_k: 0, and repetition_penalty: 1.0, generating up to 1024 new tokens.

Use Cases

This model is particularly well-suited for applications requiring text generation or understanding within a press conference scenario. Its fine-tuning suggests an optimization for generating responses, summaries, or dialogues that align with the typical structure and tone of such events.