Mikkkkoooo/qwen35-4b-private-analyst-full-corpus
Mikkkkoooo/qwen35-4b-private-analyst-full-corpus is a 4.5 billion parameter Qwen 3.5 model fine-tuned by Mikkkkoooo on a private financial research report corpus. This model specializes in generating analyst-style financial commentary and supporting private research workflows. It is optimized for internal use with sensitive financial data, offering capabilities for further fine-tuning and evaluation.
Loading preview...
Overview
This model, Mikkkkoooo/qwen35-4b-private-analyst-full-corpus, is a 4.5 billion parameter Qwen 3.5 variant fine-tuned from unsloth/Qwen3.5-4B. It was trained on a private, cleaned corpus of 23,974 financial research documents, with a sequence length of 1024 tokens over a single epoch. The training data was generated using a hybrid-review draft workflow, emphasizing its suitability for internal, private applications.
Key Capabilities
- Analyst-style financial commentary: Generates text in the style of financial analysts.
- Private research workflows: Designed for use with sensitive, proprietary financial data.
- Internal fine-tuning: Serves as a strong bootstrap artifact for further internal fine-tuning and evaluation.
Limitations and Intended Use
Due to its training on private financial documents, this model is intended for private use only, unless underlying data rights are explicitly cleared. The training set is draft-generated, not human-labeled, meaning the model should be considered a bootstrap artifact rather than a production-ready checkpoint. No public benchmarks or human evaluations are included. It is suitable for:
- Generating financial commentary.
- Supporting private research.
- Serving as a base for further internal development and evaluation.